Guest Post by Willis Eschenbach
So once again, I have donned my Don Quixote armor and continued my quest for a ~11-year sunspot-related solar signal in some surface weather dataset. My plan for the quest has been simple. It is based on the fact that all of the phenomena commonly credited with affecting the temperature, such as cosmic rays, the solar wind, changes in heliomagnetism, changes in extreme ultraviolet (EUV), or changes in total solar irradiation (TSI), all vary in phase with the sunspots. As a result, if there is no sunspot cycle visible in the terrestrial surface weather datasets, then we can assume that none of those phenomena are affecting the dataset.
To date I’ve looked for an 11-year cycle in local sea level datasets, river flow datasets, beryllium datasets, global surface temperatures, lower tropospheric temperatures, global sea levels, Nile River levels, and a bunch more surface datasets. I have looked at literally dozens of measures of the weather, and I’ve found … well … nothing.
Now, does this mean that the sunspot cycles have no effect on the weather? Absolutely not, you can’t prove a negative in any case. It just means, I haven’t found the evidence to support the claims that some weather phenomena somewhere follows the sunspots. It may still be there, and so I continue looking.
I freely admit that I entered upon this quest with a preconception of what I’d find … but not the preconception I’m often accused of having. What many people may not realize is that I thought this quest would be easy. I believed that the evidence of a sunspot-climate connection wouldn’t be hard to find. Back in 2008 I wrote:
There are strong suggestions that cloud cover is influenced by the 22-year solar Hale magnetic cycle.
And given the number of places I’d read the claim that sunspots influence the climate, I figured it would be no problem to find surface weather datasets that showed some correlation with the 11 or the 22-year sunspot cycle. But to my great surprise, I’ve not been able to come up with a single surface weather dataset of any kind which shows such effects.
In this quest, I have relied mostly on a couple of methods to determine if the sunspot cycles are present. One is the cross-correlation analysis. This shows how well correlated two time series are at a variety of time lags. It helps in distinguishing real from random results—for example, if the correlations are as great for sunspots lagging the climate as they are for sunspots leading the climate, you’ve got problems.
Note that there is an inherent problem when looking for correlations with a signal like sunspots, which has a strong cyclical nature and high autocorrelation. This is that you will get what look like real correlations when you correlate sunspots with a wide range of random data. In addition, again because of the strong cyclical nature of sunspots, whatever correlations you find you’ll also find around eleven years later. As a result, you have to include an adjustment for autocorrelation when you calculate the statistical significance of your results, or you need to do a Monte Carlo analysis of the question.
The other method that I rely on is the periodogram. This is an analysis of the underlying cycles in the data itself. For example, here are the sunspots since 1749:
Figure 1. Monthly sunspots since 1749. Pre-1947 values have been increased by 20% to bring them into line with the modern counts, as has been repeated discussed here at WUWT. It makes little difference for the analysis.
And here is the periodogram of the sunspots shown above:
Figure 2. Periodogram of monthly sunspots. A periodogram shows the strength of each of the component cycles. The units are the percentage of the range of the data (range of individual cycle divided by range of the data. Due to popular demand, the data is windowed with a Hanning window.
Note that there is a fairly broad peak with the largest value at 11.2 years, and a secondary narrow peak at 10 years. So if we find these kinds of signals in any surface weather dataset, that would be good evidence that there is a connection to either the sunspots, or to some phenomenon that varies in phase with the sunspot cycle, such as cosmic rays or TSI.
With that as prologue, let me show the latest dataset I’ve been investigating. This is US station records showing the percentage of clouds. The cloud data are from 192 locations scattered around the US. As is my custom, I started out by looking at each and every one of the 192 records in the dataset. Here is a sample of ten of them:
Figure 3. Ten of the individual cloud records in the dataset.
When I looked at that, I was surprised by the trends. To be sure, the scale is different for each panel, which emphasizes the trends. But the same results are visible in the average of the dataset.
Figure 4. Average of all of the 192 cloud datasets. Unfortunately, the data ends in 1987. The data is calculated as the cumulative sum of the means of the first differences (monthly changes) of all available datasets. Note that in this case, this first-difference method shows little variation in the period 1900-1987 from a simple mean of the data (not shown).
Now this is what I call an interesting result. I would never have guessed that the cloud cover in the US was about 10% greater in the 1980s compared with the 1920s. I do find it interesting that in response to the rapid US surface temperature rise during the 1930s and 1940s, the cloud cover acted to minimize the change by increasing quite rapidly.
However, that’s just evidence in support of my hypothesis that the clouds (and other emergent phenomena) act to minimize changes in temperature. It says nothing about whether there is a solar signal there or not. To determine that, I calculated the periodogram for the average cloud percentages shown in Figure 4.
Figure 5. Periodogram, average of US station data.
As you can see, there is very little in the way of power in the range from five to twenty-five years … no sign of a solar signal.
The situation with the individual datasets is much the same. There are a variety of signals, but again nothing in the range from five to twenty-five years exceeds about 5% of the total signal. Here is a typical set of four:
Figure 6. Periodograms showing the strength of the various cycle lengths, for four individual stations.
As you can see, there is quite a bit of variation in the individual records. Some have strong annual cycles, while others have equal-sized annual and six-month cycles. And some show only a six-month cycle, with no clear annual cycle. This is interesting, because it allows us to identify those stations which have strong thermally-driven summer thunderstorms. They show strong six-month cycles.
But to return to the sunspots, none of the stations shows any significant strength in the range from five to twenty-five years, it’s all down in the weeds, and each of them is different.
So I had to conclude once again that the solar signal is simply not visible in these cloud datasets.
Once I was done looking at the solar question, I got to thinking about the fact that the ground station cloud dataset ends in 1987. I realized that the ISCCP satellite dataset started in 1983 … giving a 53-month overlap. So I decided to see what the ISCCP dataset had to say about the US clouds. Since the ISCCP dataset is gridded, it is a matter of extracting just the land data in the area covered by the US stations. I started by looking at the two datasets during the period of the overlap:
Figure 7. Overlap of the two cloud coverage datasets, one being the average of 184 ground station records (green), and one from the ISCCP satellite data (blue). In the raw form, the two are offset by only about one percent. As a result, I have rebaselined the ground station data upwards by about one percent to match the ISCCP data during the period of the overlap.
Now, I have to say that I was very impressed by the agreement between the two datasets. I had expected to find much poorer correlation. But in fact, by and large the two datasets are in surprisingly good agreement.
As a result, I think that we are justified in splicing the two datasets together, so that we can get a long-term record of the US clouds. Here is that record.
Figure 8. Spliced data, with the early data from the ground stations and the late data from the ISCCP satellite. I used the average of the two datasets during the period of the overlap.
Now, this is a fascinating graph containing lots of unanswered question. It puts the ISCCP data (black and red) into a most interesting long-term context. When looking at just the ISPPC data it seems as if the cloud coverage was falling. But in the longer context, it can be seen that this kind of variation has been going on since about 1960.
I draw no great conclusions from this, except that overall, as the US warmed post-1930, the US cloud coverage rose as well. Go figure …
One reason I’m reluctant to draw many conclusions about this graph of the clouds is a curious one. In science, it is very difficult to demonstrate causality. However, a very smart man named Granger realized that we can demonstrate a weaker form of causality, called “Granger causality”.
The essence of the idea of Granger causality is that if having historical information about phenomenon A allows us to significantly improve our predictions of phenomenon B, then A is said to “Granger-cause” B. Makes sense to me. If knowing historical cloud cover allows us to improve our predictions of temperature, then cloud cover would be said to “Granger-cause” temperature.
However, there’s an oddity of Granger causation, which is that there are not just three, but four possible states, viz:
1) Neither variable A nor variable B Granger-causes the other, or
2) Variable A Granger-causes variable B, or
3) Variable B Granger-causes variable A, or finally
4) Variable A Granger-causes variable B AND variable B Granger-causes variable A.
So … care to guess the Granger relationship between temperature and clouds? Yep, you’re right, it’s choice 4) … each one Granger-causes the other one. Clouds affect the temperature, and temperature affects the clouds. Knowing the history of either one helps us to predict the next step of the other one. It’s just another demonstration of Murphy’s Corollary, which is that “Nature always sides with the hidden flaw.” Which definitely makes it hard to draw lots of conclusions.
In any case, that’s all the fun I’m legally allowed to have in one twenty-four-hour period. So to keep the future full of interest, let me put out a request for people to send me a link to a dataset. What dataset? Why, of course, the surface weather dataset that you think most clearly shows the sunspot-related solar influence.
A few notes about the kinds of datasets I’m looking for.
First, I’m asking what you think is the very best dataset, the dataset that actually contains the clearest evidence for the solar signal. Please don’t send me links to ten datasets. I don’t have graduate students or assistants and I do have a day job pounding nails. So please, choose your best evidence to make your best case.
Note that I’m not asking for studies, I’m asking for observational data. I’m not interested in any of the hundreds of studies out there that claim to find a solar signal in some surface climate dataset, like my favorite study. That was the one that found the solar signal … in the ring widths of one single tree in Chile. Seriously, one single tree, the study is out there, and somebody sent it to me claiming it was evidence.
Here’s the issue. Every study that I’ve looked at has had serious and often horrendous problems, and I’m sick of digging through the trash looking for a diamond. In any case, no matter what the study may claim, I’m gonna have to do the analysis myself. And to do that, I don’t need the study—all I need is the link to the dataset.
Next, please provide a link to the exact dataset itself. I don’t want to guess which dataset you are referring to, or choose from among a dozen datasets on some web page, or have to google some obscure name.
Next, please no “reanalysis data” from NCAR or NCEP or anywhere. Let me explain the problem with using what is laughably called “reanalysis data” from NCAR or NCEP in any analysis of solar effects. The “reanalysis data” is not data at all. Instead, it is the output of a computer model. Now, I’m not an opponent of computer models, I’ve both written and also used lots of computer models. The problem is that the current generation of climate models are quite linear … but nature is antilinear.
Now, for some kinds of analyses this is not a big problem, but this flaw makes model output useless for some specific applications. In particular, since the model outputs are some kind of lagged linear transformation of the inputs, in general whatever you put in the front end of a climate model will be visible in the output … and downward solar radiation flux is most assuredly an input to the reanalysis models. Since the models are linear, this means that there is a very good chance that we will find a solar signal in the output of the reanalysis models, that we’ll see a sunspot-related wiggle in the so-called “reanalysis data” … but unfortunately finding that signal in the computer output means absolutely nothing about the extremely non-linear real world. So please, if it says NCEP or NCAR or “reanalysis” anywhere, don’t bother sending it.
Finally, as a radio amateur I’m well aware of the solar effect on the ionosphere. Up there where there is one molecule per cubic ohmmeter or somesuch miniscule number, and it’s not atmosphere as we know it but a plasma of ionized gases, yes, I agree that the sun has effects. However, I can find no evidence that those effects make it down to affect the weather where we live, at the surface. Might do so, but I haven’t seen it yet. This is why I’ve asked for surface datasets, because I’ve never found the claimed 11-year cycles down here at the bottom of the gravity well.
So with those exceptions, you’re free to send any of a host of surface datasets that you think have a sunspot signal—sea level, rainfall, river flow, lake levels, temperature, atmospheric pressure, river flow and many more, the choice is up to you.
Now, there is a commenter who objects strongly whenever I ask people to send me their best evidence. He claims I’m asking him to do the research that I should be doing. But all I’m asking for is a dataset—I am the person who has to do the research. The problem is, there are hundreds and hundreds of datasets out there, and my time is short. I don’t have time to analyze them all. So I’m asking for your assistance in winnowing the wheat from the chaff. If you don’t like my asking for assistance in this process, please, just ignore the post. Raising the same objections once again won’t burnish your reputation.
In addition, I see this as a chance for you guys to definitively show everyone that there actually is a sunspot-related signal that makes it down here to the surface where we live. So I invite you—bring out your best dataset. I’ve just looked at the US cloud record and found nothing … your turn.
Regarding the clouds, I walked this evening with the gorgeous ex-fiancee along the western edge of Bodega Bay. The cold front that brought us welcome rain earlier in the day was just past, and so we got to look east across the Bay at the back side of the cold front. Mostly it was kind of a wiggly stratus, an unusual formation. Out to the west, there was clear sky, and the late evening sun from behind us lit up the clouds above and the hills below. Looking at that blue, clean-washed piece of sky, I was reminded of the scientifically accurate words of the poet about clouds:
I am the daughter of Earth and Water,
And the nursling of the Sky;
I pass through the pores of the ocean and shores;
I change, but I cannot die.
For after the rain when with never a stain
The pavilion of Heaven is bare,
And the winds and sunbeams with their convex gleams
Build up the blue dome of air,
I silently laugh at my own cenotaph,
And out of the caverns of rain,
Like a child from the womb, like a ghost from the tomb,
I arise and unbuild it again.
Best regards to everyone, keep on unbuilding,
w.
MY REQUEST: If you disagree with someone (highly unlikely, I know, but work with me here), please QUOTE THE EXACT WORDS YOU DISAGREE WITH. That will let everyone know just what you think is incorrect.
MY POSITION: Note that I am NOT saying that the sun doesn’t affect the earth. It does affect the earth in a host of ways, and we can see that every dawn.
What I AM saying is that I have never found a significant sunspot-related ~11 year signal in any surface weather-related dataset. Doesn’t mean it’s not there somewhere, just means I haven’t found it yet. If you think it’s there … this is your chance to identify the dataset which contains the signal.
DATA: the ISCCP D2 cloud coverage data is here, the US ground station cloud coverage data is here, and the sunspot data (unadjusted) is here.
CATALOG: all of the ISCCP data is indexed here. Took me a while to find it, because it’s on the NOAA website.
CODE: R computer code, along with some of the collated data, is here as a zipped folder. The code is all there. However, I’d describe it as user-aggressive … and like all of my code, it’s not written to run from top to bottom. It also contains a lot of other things that don’t make it into the final report, various analyses and graphs. I believe that all of the code and functions are there to produce all of the graphics and analyses above. If not, let me know.
Why is one looking for a 11 year cycle?.
No one seriously suggests that temperature fluctuates on a 11 year basis, or that sea level rises and falls (or the rate of chabge thereof varies) on a 11 year basis.
The fact that one cannot find some correlation in some data set (which in any event would not establish causation) with an 11 year cycle sheds all but no light on whether solar may be a major driver of earth’s climate. The absence of such correlation, certainly does not establish that solar is not a significant player.
richard verney November 1, 2014 at 1:58 am
Thanks, Richard. Sadly, there are a host of people who seriously suggest exactly what you say. For someone seriously suggesting that “sea level rises and falls on an 11-year basis”, see here. And for someone saying “temperature fluctuates on a 11 year basis”, see here.
As to why I’m looking for an 11-year cycle, it is because all of the relevant phenomena, such as galactic cosmic rays or TSI, vary on an ~11-year cycle. So if one or more of those phenomena are actually affecting the weather, we’d expect to find an 11-year cycle in some weather dataset.
Finally, you say not finding an ~11-year cycle “sheds all but no light on whether solar may be a major driver of earth’s climate”. I’m not sure what you mean by a “major driver”, but regardless, I’d say that the sun is a major climate driver.
However, that sheds no light on whether cosmic rays affect the weather … for that, we need to look for the 11 year cycle, despite the fact that the sun is indeed a major climate driver.
In other words, you are right that not finding the ~11-year cycle means nothing about whether the sun is a major climate driver … but it means a whole lot about whether cosmic rays are a major climate driver.
All the best to you,
w.
Energy storage systems tend to act as low pass filters to the input. ‘Nuff said.
ShrNfr November 1, 2014 at 6:06 am Edit
Thanks, ShrNfr, but without an actual example and the math to back up your claim you’re just blowing wind …
w.
I agree with Richard. Willis, I think your premise is that the system is simple, so that variations in the input signal should be visible in the output. However, the system is extremely complex. Also, your approach is mathematical (like CA), instead of scientific.
For example, what if we were looking at datasets for Lake Ontario’s water level, and found no correlation to rain storms in the upper great lakes region? The system is too complex for that to show up.
Viking, you and others seem to think that I invented the idea that there are ~11-year sunspot cycles in the climate. In fact, the idea goes all the way back to the claims (since refuted) by William Herschel that there were sunspot cycles visible in wheat prices. So I’m just trying to refute that idea, but I make no assumptions about whether the system is simple or not.
If you truly want me to consider that example, please provide links to the datasets that support your claim.
w.
Willis, it’s a good exercise in the scientific method to attempt to falsify every idea, especially our own. So, good on you.
However, you may have missed my point. My point is that when systems are complex enough, input variations may not be detectable in the output. Therefore, your search for this type of correlation falsifies nothing.
I presented the great lakes system not as a data set related to the 11 year sunspot cycle, but as a system too complex to see variations in upper great lakes region rain storms (input) in the Lake Ontario water level (output).
Willis
It’s very interesting that nothing is detectable anywhere that connects the sunspot cycle to climate. Although when searching for exoplanets one of the things looked for is if they are in a habitable zone. It just suggests to me that two things are preventing establishing a link between the Sun, in general, and Climate (other than there isn’t one which isn’t likely)
1. Not enough data
2 Looking in the wrong place.
I prefer to look at periods of not less than 30 years or about three solar cycles because the single cycle variations swamp the longer term trends. and to use data from the Earthshine project because that deals with global albedo and not just regional cloudiness changes.
Thus far the data from Earthshine is very limited historically but I do note that the Earthshine trend changed around 2000 as did the global temperature trend.
I think it will be found in due course that an active sun reduces global cloudiness, allows more energy into the oceans, skews ENSO in favour of El Nino relative to La Nina and causes warming overall.
The opposite for a quiet sun.
As for the mechanism I have set that out elsewhere.
Well, good luck with your tilling at 11 year windmills. 0.8C in 150 years? No way you are going to find it like that.
As I have said many times before – “If you don’t understand how the sun heats the oceans, you cannot understand climate”.
Given that you have shown no understanding of the multiple SW selective surface effects of liquid water, including –
1. SWIR/SW/UV absorptivity higher than IR emissivity.
2. SW/UV absorption at depth, not surface.
3. Intermittent illumination.
4. Internal convection.
– Just how do you expect to claim no discernible solar effect? Dr. S has worked it out. That’s why he is trying to stamp not just TSI flat, but UV variance as well.
Everybody is watching. Many are laughing 😉
(seriously, if you cannot understand how the sun heats the oceans, how can you determine the effect of cloud cover variance?)
Konrad. November 1, 2014 at 2:40 am
Thanks, Konrad, but I have to ask—was there some part of QUOTE THE EXACT WORDS YOU DISAGREE WITH that escaped you? Because I can repeat the explanation if you wish.
If you can quote me saying somewhere that SW/UV is NOT absorbed at depth, for example, we could understand what your point 2 is about.
Or take number 4, that I don’t understand the internal convection of the ocean … see here for a discussion of that exact question, which I have also described elsewhere. If there is some part of that you think is wrong, please QUOTE THE EXACT WORDS that you disagree with.
A scattergun attack like the one you’ve delivered above is meaningless, unpleasant, and detrimental to both your reputation and further discussion. No one knows what you are talking about.
Finally, I note that in all of that you have NOT provided any dataset that shows the ~11-year solar cycle.
w.
PS—as far as I know, I’ve never commented on the ratio of IR emissivity and SW absorptivity. Let me remedy that. IR emissivity of the ocean is about 0.96.
SW absorptivity is 1 – albedo. The albedo of the ocean varies from about 0.13 (rough water, low sun) to .02 (calm water, low sun). Typical mid-day conditions (relatively calm, sun within two hours of overhead) have an albedo of somewhere around 0.06. Data from my bible, Geiger’s “The Climate Near The Ground”.
So I’d say that SW absorptivity is about the same as IR emissivity, maybe a bit less. I’ll take a look at the CERES data, I should be able to figure the average albedo of the ocean … hang on … OK, CERES data puts the average ocean albedo at 0.06. Dang, go figure, my back-of-the-envelope calculation was right on.
This means that for the ocean, IR emissivity is 0.96, and SW absorptivity is 0.94.
As a result, it appears that your claim that SWIR/SW/UV absorptivity is higher than IR emissivity is not true. And more to the point … so what? I fail to see how my understanding of that question is supposed to disqualify me from looking for an ~11-year cycle.
Again, please, if you disagree with my analysis of absorptivity and emissivity, please QUOTE MY EXACT WORDS. It saves heaps of time, and gives substance to your arguments.
Willis Eschenbach
November 1, 2014 at 3:33 am
////////////////////////////////////////////////
”Thanks, Konrad, but I have to ask—was there some part of QUOTE THE EXACT WORDS YOU DISAGREE WITH that escaped you? Because I can repeat the explanation if you wish.”
I can cut and paste the exact words if you desire. Remember the “Usoskin Et Al. Discover A New Class of Sunspots” thread? The Internet does, forever.
It matters not if you understand that UV/SW is absorbed at depth or whether you understand the diurnal overturning cycle. You have shown you don’t understand the effect of this on average ocean surface temps. You are still stuck in oceans=near blackbody nonsense land. The oceans are convecting, SW translucent and intermittently illuminated. Here are the 5 rules –
For SW translucent / IR opaque (material A) compared to SW opaque / IR opaque (material B) with both materials having equal IR emissivity and total watts for both constant or intermittent SW illumination being equal, the results of empirical experiment are clear –
1. If materials are solid, constant SW illumination will result in close surface temps for A & B with average temp of A higher than B
2. If materials are solid, intermittent SW illumination will result in surface temps for A higher than B, with average temp of A also higher than B.
3. If materials are liquid and convect, constant SW illumination will result in surface temps for A higher than B, with average temp of A higher than B.
4. If materials are liquid and convect, intermittent SW illumination will result in higher temperature differential (both surface and average) between A & B than condition 3.
5. If materials are liquid and convect, intermittently SW illuminated and deeper than condition 4, temperature differential between A & B will be greater again than condition 4.
Those 5 rules are verified by empirical experiment. Resistance is useless. You will be assimilated.
And as to the asymmetry between SW absorptivity and IR emissivity? Your referencing in situ measurements don’t count. Those are just measurements of apparent (not effective) emissivity within the Hohlrumn of the atmosphere. The simplest empirical experiments –
http://i61.tinypic.com/24ozslk.jpg
– show apparent emissivity dropping below 0.8 at 15 degree viewing angle when background temp is dropped below -40C.
Smarter people than me have measured BRDF using a modulated source and a synchronous detector for still water. Answer? Emissivity dropping below 0.05 beyond 70 degrees from vertical.
E=0.96 for water? Apparent emissivity (within the atmospheric Hohlrumn) maybe. Effective emissivity? Don’t make those watching laugh any harder, they may cough up their spleens!
Willis, you were right about the “Iris” or “cloud thermostat”, your “steel greenhouse” works (only with vacuum), I know, I built it. But you are just going to have to accept that “lukewarmers” were just as scientifically incorrect as CAGW believers.
Am I the bad guy Willis? How many times have I stated that I wanted you to be the one that presented the right answer? Face facts, you were given the clues, but your ego was the obstacle.
That is TILTING at windmills, an English idiom (from jousting) which means attacking imaginary enemies.
Are sunspots just a figment of the imagination?
While “tilting” indeed does come from jousting, the phrase “tilting at windmills” derives from Cervantes’ novel “Don Quixote”, referenced by Willis.
Dennis,
Tilting at windmills is one thing. Re-ploughing “the sun does nothing” soil (tilling) is quite another 😉
Willis has not yet descended to the depths of Dr. S, but he is getting dangerously close……
Konrad. November 2, 2014 at 3:24 am Edit
Konrad, this is why I say QUOTE THE EXACT WORDS YOU DISAGREE WITH. My exact words in this were not “the sun does nothing”, as you falsely claim. My words were:
I put that in there specifically because I knew that some jerkwagon like you would try to misrepresent my position … and as I foresaw, you’ve tried just exactly that.
I never said “the sun does nothing”, that’s just your ugliness surfacing. Your attempt to put words in my mouth, and then attack me for what you falsely claim I said, is pathetic. Either QUOTE MY WORDS or go play somewhere else. Your unpleasant, uncited, and mendacious attacks are getting old.
w.
”I never said “the sun does nothing”, that’s just your ugliness surfacing. Your attempt to put words in my mouth, and then attack me for what you falsely claim I said, is pathetic.”
Willis,
I am most assuredly not attempting to put words in your mouth. I am therefore not attacking you for any false claims about what you said. My concern is your approach to the question of solar influence. While you do say –
”Note that I am NOT saying that the sun doesn’t affect the earth”
You also say –
” It is based on the fact that all of the phenomena commonly credited with affecting the temperature, such as cosmic rays, the solar wind, changes in heliomagnetism, changes in extreme ultraviolet (EUV), or changes in total solar irradiation (TSI), all vary in phase with the sunspots. As a result, if there is no sunspot cycle visible in the terrestrial surface weather datasets, then we can assume that none of those phenomena are affecting the dataset.”
What I and many others are saying is that just because there is no discernible 11 year cycle in current data sets, it does not necessarily follow that the phenomena listed (and the addition of TSI component variation in UV bands other than EUV) are not having longer term climate effects. In fact, given the way the sun heats the oceans, and the complexity of ocean circulation, I would consider it unlikely that a clear 11 year signal would be manifest.
This means I am not so much objecting to your words, but actions. This is one of a string of articles on not being able to find an 11 year cycle in messy data. Why do you persist with these? The lack of a 11 year signal in no way rules out a long term solar influence. Further, because the most plausible mechanisms (cloud cover variation and TSI component variation) effect energy accumulation below the ocean thermocline, looking for an 11 year signal would be looking in the wrong place.
Such a signal may in time develop in the ARGO data, but currently we have insufficient data. If two extra sensors were included on the new “ARGO deep” probes, we may have some chance of early detection before the full run of SC25. Those would be optical ocean turbidity sensor and multi frequency solar penetration sensor.
All that can be said at this time is we have insufficient data to rule out a long term solar influence on climate. No amount of articles on missing 11 year signals can change this, and this raises the question – what is the point of these no 11 year signal articles?
If you recorded temperatures above a storage heater that was turned on for one hour in every six and then tried to detect the six hourly cycle you would probably fail. The decline in heat output during the intervening 5 hours would probably be swamped by other factors such as outside temperatures, the number of people in the room etc.
Trying to detect an eleven year signal in earth temperatures, which is a much more chaotic system, seems, to me, to be a similar task as the sea acts as the earth’s storage heater. We know that at the end of an ice age there is an average lag of 800years before CO2 levels rise. This is because it takes that amount of time to warm the seas enough to start releasing gasses.
If on the other hand the storage heater was switched alternately on and off for a two week period the difference in temperatures would be easy to detect.
I suggest that you look for a correlation between sunspot activity and temperatures when there is a much longer period (as during the Maunder Minimum), but then this link is already well established.
Excellent point.
So that would lead me to believe that the small change in solar measures over the recent rise and pause of surface temperatures can be disregarded (essentially set at 0), because any causality under the recent past 60 years is buried in the weeds of Earth and its phenomenal ability to ignore such small solar changes and do its own thing.
There are at least two regular commentators here who ascribe to that very theory. Ignore the solar connection until a certain level of extreme solar variation is present. I dare say this premise is also ascribed to and can be found in research articles penned by several kingpin climate scientists: Only under extreme solar variations do we find possible correlations and with unsubstantiated mechanisms (some provide suggested mechanisms).
Pamela, there is a difference between a) system ignoring input variation and b) system being complex enough to prevent a mathematical correlation between output and sinusoidal input.
PaulM November 1, 2014 at 2:49 am
Oh, please. That’s just an empty claim. In many cases, yes, you would definitely be able to detect the signal.
Hey, I’m not the one making the claim that there are ~11-year cycles in the climate. There are literally dozens of people making that claim. I’m just the poor fool looking for the evidence for that claim.
No, that link is no better established than the claims of the 11-year cycles. Yes, the Maunder Minimum (MM) occurred during the Little Ice Age (LIA) … but the LIA started before the MM and continued after the MM, so the MM obviously didn’t cause the LIA. I have never seen a single scrap of evidence that the MM was colder than the periods immediately before or after. If you have such evidence, now would be a good time to present it.
Until you come up with said evidence, you might enjoy reading this …
w.
“I have never seen a single scrap of evidence that the MM was colder than the periods immediately before or after. If you have such evidence, now would be a good time to present it.”
I was amazed that you could not see it for yourself, but to ignore it after it was pointed out to you several times is inexcusable:
http://wattsupwiththat.com/2014/06/23/maunder-and-dalton-sunspot-minima/#comment-1668036
Ulric, the problem is that we have so little data for the time of the Maunder Minimum. About the best we have might be the “Winter Severity Index” of the famous H. H. Lamb. If you could point out the Maunder and Dalton minima on this, I’d be obliged …
… did you get it right? Because they are here:
…
…
…
Aw, heck, I got my dates mixed up. Here’s the actual data:
As you can see, the dates of the various “minima” are not immediately obvious, because … well, because they seem to make little impression on the temperature.
Ulric, I’d said:
Regarding the graphic you directed me to, that was this one:
Unfortunately, that graph is uninformative as to whether the temperature was warmer or colder before the MM, since the data starts after the start of the MM. It does, however, show the Dalton minimum. According to the CET (Central England Temperature), England warmed over the period of the Dalton Minimum.
The problem is that there is a lack of actual data to back up the claims of cold times during the various minima.
There is a deeper problem, however. This is that the changes in sunspots/TSI/GCR/whatever between say the Dalton Minimum and the period on either side is much smaller than the 11-year swings in the data. Why on earth should the planet be sensitive to and respond quickly to daily and annual swings in the sun’s strength, and be insensitive to 11-year swings in the sun’s strength … and despite all of that, be very sensitive to even tinier swings in the long-term activity of the sun?
That’s the conundrum I don’t understand. The temperature of the earth’s surface responds very quickly to small changes in solar input. The thinnnest hazy cloud comes over the sun and the surface cools right away. But the changes from the 11-yer cycle are not visible. I say that’s because they are so small that they get lost in the weeds … but the changes to/from the Dalton Minimum are even smaller than the 11-year cycles.
So how do you explain the fact that we can’t find the 11-year cycles, and yet you claim the much weaker cycles to/from the solar minima make some easily visible change in the temperature? That’s my problem, I can’t figure out physically how that would happen.
w.
Willis,
You don’t have to rely on the Winter Severity Index. Others & I have repeatedly linked here to the reconstructed CET going back to AD 800.
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&ved=0CC0QFjAD&url=http%3A%2F%2Fwww.climateaudit.info%2Fpdf%2Fothers%2Flamb.ppp.1965.pdf&ei=7HVWVIaBMJeaoQTHmIL4DQ&usg=AFQjCNGMR8vloZcVY6Gt2RF9TUw0jvSK4w&sig2=6UEVn4JJDKaeMJxE6W4s-A&bvm=bv.78677474,d.cGU
Table II (PREVAILING TEMPERATURES (°C) IN CENTRAL ENGLAND, Annual Averages) clearly shows the effect of minima. These 50-year averages don’t correspond directly to the usual dates for the minima, & the longer periods include rebounds from low temperature decades. But the effect of the minima is still visible, despite the imprecise match, with the periods straddling warm & cold intervals or including both.
Tony could provide the annual data for greater resolution, if you’re interested, since the effect shows up more dramatically using shorter periods.
800-1000 9.2
1000-1100 9.4
1100-1150 9.6
1150-1200 10.2
1200-1250 10.1
1250-1300 10.2
1300-1350 9.8 Wolf
1350-1400 9.5
1400-1450 9.1
1450-1500 9.0 Spörer
1500-1550 9.3
1550-1600 8.8
1600-1650 8.8
1650-1700 8.7 Maunder
1700-1750 9.24
1750-1800 9.06 Dalton
1800-1850 9.12
1850-1900 9.12
1900-1950 9.41
Here is Tony’s own extension back to 1538:
http://wattsupwiththat.com/2013/08/16/historic-variations-in-temperature-number-four-the-hockey-stick/
milodonharlani November 2, 2014 at 11:37 am
Say what? That link goes to Lamb’s work that includes the very Winter Severity Index that I listed.
w.
Willis said:
“Unfortunately, that graph is uninformative as to whether the temperature was warmer or colder before the MM, since the data starts after the start of the MM. It does, however, show the Dalton minimum. According to the CET (Central England Temperature), England warmed over the period of the Dalton Minimum.”
The graph without question shows the Dalton Minimum as one of the three coldest periods through CET, the other two being during the Maunder and Gleissberg minima, and it also shows that the MM was *colder than any period after it*.
The weaker sunspot cycles of the MM began around 1650, and we have better records than than Lambs winter index for the 1650’s. Apart from two noted cold winters, it was generally warm, and very warm in the first 5 years:
http://wattsupwiththat.files.wordpress.com/2011/09/weather1.pdf
Then CET takes over from 1659.
“There is a deeper problem, however. This is that the changes in sunspots/TSI/GCR/whatever between say the Dalton Minimum and the period on either side is much smaller than the 11-year swings in the data.”
Not really so for SSN, SC4 SSN = >140, SC’s 5&6 SSN = 145. The most curious thing though is that the period 1836-1845 in CET is as cold as the coldest part of Dalton; 1807-1817. The most pertinent clue to the coldest part of Dalton that I can see, is the dearth of Aurora that Silverman notes (page 11):
http://www.leif.org/EOS/92RG01571-Aurorae.pdf
“Why on earth should the planet be sensitive to and respond quickly to daily and annual swings in the sun’s strength, and be insensitive to 11-year swings in the sun’s strength … and despite all of that, be very sensitive to even tinier swings in the long-term activity of the sun?”
I would not say an El Nino response to slow solar wind is insensitive. The changes in the aa index in the last 100 years were not tiny:
http://www.ngdc.noaa.gov/stp/geomag/aastar.html
p.s. I don’t agree with the popular dating of the Spörer Minimum, there should be two minima there, one from the 1430’s, and a second from the 1550’s, as confirmed by CET reconstructions and proxies.
Some of my text got lost there, it should have read:
SC’s 5&6 SSN = 145.
Same problem, I’ll spell it this time. SC’s 5&6 SSN was less than 50, and in SC8 SSN was greater than 145.
Good point! I still think that Svensmark’s theory is valid, but Lief will chop my head off for saying that!
http://iopscience.iop.org/1748-9326/7/4/044004/pdf/1748-9326_7_4_044004.pdf
Persistent solar signatures in cloud cover: spatial and temporal analysis
Voiculescu & Usoskin 2012
You forgot to provide a direct link to the data set they used in that study.
Pamela, I didn’t forget. There is a link in the link and even if there weren’t a link he has the link already.
Okay. I see that you have not quite understood Willis’ directions. But I’ll bite. I linked on the link and got this:
http://isccp.giss.nasa.gov/products/onlineData.html
Lots of choices here. Now tell us which one is the data set used in the study you cite so we can all know that that one is the data you think best to use.
Personally I’d go with the D2 data since that’s what they specified. If you doubt you can find the right data perhaps writing the author at the email address he provided for instructions would help you. I would except I have no intention of doing anything with the data.
Thanks, Steven. To begin with, it’s not at all clear exactly which dataset they used. You blithely say that “Personally I’d go with the D2 data since that’s what they specified.” … but it appears that you are totally unaware that there are no less than nineteen D2 datasets, viz: MnCldAmt.nc, MnCldAmt-St-Liq.nc, MnCldAmt-St-Ice.nc, MnCldAmt-Sc-Liq.nc, MnCldAmt-Sc-Ice.nc, MnCldAmt-Ns-Liq.nc, MnCldAmt-Ns-Ice.nc, MnCldAmt-Mid.nc, MnCldAmt-Low.nc, MnCldAmt-Hi.nc, MnCldAmt-Cu-Liq.nc, MnCldAmt-Cu-Ice.nc, MnCldAmt-Cs.nc, MnCldAmt-Ci.nc, MnCldAmt-Cb.nc, MnCldAmt-As-Liq.nc, MnCldAmt-As-Ice.nc, MnCldAmt-Ac-Liq.nc, MnCldAmt-Ac-Ice.nc.
All of those are D2 datasets … care to tell us which ones they used, and let us know how you determined which ones they used?
In any case, that is a classic example of a data dredge. Let’s look at one of their results:
ORIGINAL CAPTION: Correlation map between high cloud cover and CRII (negative correlation: blue dots, positive correlation; red dots, 90% confidence) and average high cloud cover for 1984–2009.
Now that all looks very impressive … until you consider a few things. First, their threshold for statistical significance is 90% … which means that with random data, we should find “significant” results 10% of the time.
But wait, it gets worse. That 10% would be what we would find if we looked at one single dataset. But according to them, their data dredge has extended to no less than three target datsets (some kind of low, middle, and high clouds) and two putative forcings (cosmic rays and ultraviolet). With that many datasets and forcings, you are almost guaranteed to find a plethora of results at the 90% level.
Let me give you an example. If you pull out a coin and flip it six times, the odds of getting six heads is one in 2^6, or one in 128. That would be a statistically significant result, it might indicate a weighted coin.
But suppose you did the six-flip sequence a hundred times. Somewhere in there you are almost guaranteed to get six heads … but so what?
Similarly, when they look at 2,500 gridcells in each of three datasets and they compare them to two different forcings, somewhere in there you are almost guaranteed to get something like Figure 1 … again, so what?
But wait, it gets worse. These naifs seem completely unaware of the fact that when you use autocorrelated data, like say the ISCCP dataset and the cosmic rays dataset and the ultraviolet dataset, you find many more “significant” correlations than you would find with random data. As a result, you need to allow for autocorrelation
But wait, it gets worse. Even though the dataset only covers a pathetic two solar cycles, the authors say:
They state that, but they never seem curious enough to wonder why they are getting spurious correlations.
But wait, it gets worse. For unknown reasons, they have downsampled the ISCCP data from the original 2.5°x2.5° grid to a 5°x5° grid, which increases the spatial autocorrelation. They make no mention of why they have done this or what effect it has on the statistics.
Short answer? Sorry, Steven, but that study is garbage. When you mine that much data and you ignore the issue of repeated sampling and you don’t deal with autocorrelation, you are guaranteed to find the kind of results they glowingly report … but unfortunately, that means nothing.
w.
Willis, You seem fairly confident in your rebuttal. Would you object to my emailing your response to the author and invite him here to defend his work? I think it would be interesting if we could get him here since both he and his coauthor have several solar studies published.
Steven, please do invite the authors to defend their work. I would be very interested in their comments.
Best of luck in your mission,
w.
Ok, I’ll see if I can get a response.
Willis, actually I’m going to hold off for now. The paper states that the best correlations were with low cloud cover and it was obvious from looking at the correlation maps that was the case. Right away attacking what appears weak instead of what appears strong makes you appear biased. I’ll take a look at some of the other issues you brought up even though statistical processes aren’t really something that interests me. I just assume that people reviewing the paper can pick out things that only take a few minutes to pick out but maybe this isn’t the case. As far as the data sets go, you are right I didn’t know there were 19. I didn’t really expect there to be only one since there were different cloud types but 19 was a bit of a suprise.
http://journals.ametsoc.org/doi/pdf/10.1175/1520-0442(1997)010%3C2147%3AAMTETS%3E2.0.CO%3B2
They state they used methods for plotting their maps that they used in the 2007 paper. In that paper they processed the data for autocorrelation using the method from the Ebisuzaki 1997 paper. A bit egotistical to think everyone is familiar with the 2007 paper perhaps but maybe those in the field are.
So far you aren’t batting too well. I think finding out why they used the grid size they did would require asking the author. Feel free.
Steven, why would serious scientists respond to questions which are posed in such a nasty, disrespectful manner? I have worked with many actual scientists, and I can say for certain that they would not deem criticism of this level worthy of response.
steven November 2, 2014 at 4:04 am Edit
Actually, they didn’t say they used methods from the 2007 paper. They said they used methods SIMILAR to the 2007 paper. This is why quoting the words you are talking about is so important. What they actually said was:
A “similar manner” could mean anything. Perhaps it means that they did adjust for autocorrelation … but perhaps it just means that they used the same mapping methods. You are starting to understand why we need the code as used and the data as used to understand a study. Saying they did something in “a similar manner” means nothing.
steven November 2, 2014 at 4:06 am Edit
Not batting too well? You have not been able to falsify a single one of my claims about the paper. You certainly haven’t shown that they adjusted for autocorrelation. You haven’t shown that they are aware of the various statistical problems I listed related to data dredging. Where am I not “batting too well”?
As to “asking the author”, after a number of refusals over a number of years, I’ve simply given that nonsense up. I tried being nice, and it got me nowhere. The most common response is just silence … which has the (perhaps intended) effect of preventing me from doing anything for a number of weeks. Sorry, but I no longer go down that path.
More to the point, this is the 21st century. Even the journals are starting to require data as used and code as used in order to avoid all of those problems. The main issue is that no matter how well you can describe in English what you THINK your computer program is doing, the program may well be (and indeed often is) doing something entirely different. They may have adjusted for temporal autocorrelation, for example, but not for spatial autocorrelation. Without the code, we’ll never know.
But heck, Steven, you seem to think it will work … so how about you write to the author and ask them to send you their code and data as used. I’m looking forward to seeing how that plays out … although if (as I suspect) they read WUWT, or at least this thread, they may give it to you just to spite me. In any case, let us know what results you get. Me, as I said, I don’t enter that maze any more. Far too often, there’s no cheese at the end …
w.
Willis, no you aren’t batting too well. It is clear they did correct for autocorrelation regardless of if they felt the need to spell it out for the reader. It is clear you picked a map that you felt would be easy to critique when there were other maps available which showed a great deal more correlation. Tell me why comparing 6 pairings for possible correlations in the same paper is worse than using the same data for 6 different papers. If you can’t that’s 3 and you have struck out. I agree with you on code so we have no argument there. Willis, I would have written him but I decided just sending your comments would reflect poorly on me. Seriously.
steven November 2, 2014 at 1:28 pm Edit
Since you are so sure they did it, perhaps you can tell us whether they adjusted just for temporal autocorrelation, or just for spatial autocorrelation, or both?
In fact, you have ASSUMED that they did it, but the entire paper doesn’t contain anything about autocorrelation … only that it is “similar to” a previous paper. And if they did adjust, we still don’t know how they adjusted You are free to interpret “similar to” as you wish, and you have done so … but that doesn’t provide any evidence. That’s why we need the code.
In fact, it was the first map I came to, and I stopped there because it was ludicrous. Once again, you have ASSUMED that you knew what I was doing, when you had no clue. Your assumption is wrong.
In any case, whether specially selected or picked because it was first, so what? It is a figure in the paper that, as you mention, is “easy to critique”. And why is it easy to critique? Because of all of the numerous problems with it that I mentioned. How it was selected is immaterial, it’s still junk.
Because they only show the successful pairings, obviously. Otherwise, they’d have to write an entire paper about a map where nothing is happening …
If you agree about the code, then why are you defending the autocorrelation question? The only reason it even comes up is because the code isn’t archived.
Finally, you said upthread that you were going to invite the authors to defend their work. Then you walked away from what you had agreed to do, saying:
Say what? Of course you start out by attacking what appears weak. You don’t attack the parts that are robust and solid, why would you? Look, my task here is to point out the flaws in the paper … but when I do point out the flaws you say oh, willis, you can’t focus on that part of the paper, that part is weak and has flaws, you should focus your search for flaws on the parts that aren’t weak and don’t have flaws …
Words fail me.
Regardless, now you say:
Seriously? Sorry, not buying that reason any more than the previous one. You don’t have to send my comments to him, why would you believe you had to do that? Nor did you say you were going to do that. You said you were going to invite him to defend his work. How inviting him to defend his work would reflect badly on you is a mystery to me.
Frankly, I don’t know why you’ve chosen not to invite the authors to defend their work, nor do I particularly care, but those excuses just don’t hold water.
Regards,
w.
Willis, strike 4 unless you can explain why a paper that is testing for these 6 possible correlations is as weak as the weakest correlation. I think we are done.
steven November 2, 2014 at 7:37 pm Edit
Steven, leaving aside the fact that you haven’t shown that I’ve made any strikes, the issue is the statistics of repeated testing. Here’s the explanation I gave above.
I can see that you don’t understand that, but I’m not sure I can explain it any more clearly. However, I’ll try.
The issue is, if you keep looking in enough places, eventually you will find something that LOOKS significant. As I mentioned, if you flip six coins one time, six heads is a very unusual outcome.
But if you flip six coins a hundred times, somewhere in there you are almost guaranteed to get six heads … which means that finding six heads in that test is MEANINGLESS. In fact, with a hundred flips of six coins, your odds of getting six heads on one of the flips is about 80% … so it would be unusual if you didn’t find six heads.
The same thing is true when looking at six datasets. If you look at one dataset, and you find a result that exceeds your threshold of the 90% level (as in this study), you have a one-in-ten chance that it’s just a random result, what’s called a “false positive”.
But if you repeat that same investigation on six different datasets, just like with the coins, your odds of finding an “unusual” result go way up. How far up? Well, they go up to (1 – 90% to the sixth power), which is (1 – .96), which is 0.47 … so when you look at six datasets instead of one, your chances of finding a false positive are almost 50/50. In other words, with six datasets, finding a result at a 90% level means nothing, there’s a 50% chance of that happening.
And that, my friend, is why looking for a result in six datasets is statistically very different than just looking in one dataset.
I must say, given the level of your understanding of these basic statistical concepts, your arrogance is somewhat surprising …
w.
There’s a fun discussion of the odds of strings of 6 consecutive heads or tails in 200 flips here. They end up figuring a 96.5% probability of this using a technique I didn’t recognize.
Willis,
Strike 4: Your attacking a weak correlation instead of a stronger correlation. CR and the correlation between high clouds is possibly, and quite likely, independent of UV correlation with low clouds. If you to critique the correlations you would pick out the strongest one not the weakest. JUst like if you had 6 different papers. You wouldn’t say this paper shows a really poor correlation between CR and high clouds so that paper showing a correlation between UV and low clouds is also poor.
Strike 3: You say a data dredge produced a map with lots of dots. I am blind and missed that map. Give me a figure number. I remember seeing the 6 that represented each of the possible correlations they were testing for.
Strike 2: It doesn’t matter if you stopped as soon as you saw something you could critique or if you looked and then picked it out. Your bias either made you stop or made you select. I’m not sure how you could start a critique of a paper without even taking the time to look at the pretty pictures first but if you say you did that’s fine doesn’t change my diagnosis of your actions.
Strike 1: You assumed they didn’t do their statistics properly because you didn’t see where it had been done. This was a no lose situation for you. You could say I don’t see where they corrected for autocorrelation and the paper isn’t likely to be correct unless they did. Instead you said they didn’t and were rather insulting about it. Now you have made it a win lose situation. Either you are right or you look foolish. I have made a declarative statement that they did. It is every bit as supported as your statement they didn’t. When I start making disparaging remarks along with my assertions we will be on the same level field.
You can make the 11 year cycle go away by averaging it – for example, if you assume a 400 sample average (roughly 30 years of smoothing (plausibly caused the by thermal inertia of the oceans?), the 11 year cycle is barely a ripple.
http://woodfortrees.org/plot/sidc-ssn/mean:400/plot/sidc-ssn
You can also play other games with the sunspot count. For example, if you subtract 40 from the sunspot count, then integrate it (so sunspot becomes a proxy for rate of heating or cooling, rather than a direct heating or cooling proxy – with less than 40 assumed to correspond to cooling), you get a graph which looks a lot like the instrumental temperature record.
http://woodfortrees.org/plot/hadcrut4gl/from:1850/mean:50/normalise/plot/sidc-ssn/from:1850/mean:50/offset:-40/integral/normalise
I don’t know whether this means anything or is just mathturbation, but I think its early to write off a solar influence – though I agree the solar argument would be a lot more compelling if there was a detectable 11 year cycle.
BTW have you looked at Willie Soon’s research into solar vs temperature? I remember a claim a while ago by Soon that he had found a correlation between solar activity and daytime maximum temperature.
http://quadrant.org.au/opinion/doomed-planet/2013/03/changing-sun-changing-climate/
The whole temperature record is hourly averages into daily average into station averages into monthly averages and then averaged with nearby stations to then on to national and world averages. Filters of proven and unproven worth are applied to the averages to get a huge pile of average data that has lost all meaning.
The WFT normalized curves are amusing. Wiggle-matching, of course. There seem to be sinusoidal components which are in opposite phase to each other…except when they aren’t. The periodicities are different, as near as I can tell. Take another look and see if anything will isolate the sine functions.
Note that sunspots are not evenly distributed between solar hemispheres. Perhaps an analysis taking this into account might be more productive. http://www.sidc.be/silso/datafiles. The 40 figure might need adjustment, or it might need to be tossed out altogether.
Monthly Data:
http://www.sidc.be/silso/DATA/monssnns.dat
Daily Data:
http://www.sidc.be/silso/DATA/ISSN_D_hem.txt
Eric Worrall November 1, 2014 at 2:52 am Edit
Wills, some years ago I wrote a post about the temperatures in Echuca, and how there was no global warming at Echuca. I also noted that the peak maximums appeared to coincide with the odd solar cycles. You poored rain down on this, in your usual manner. Maybe it is worthwhile looking at that maximum data again, along with the ENSO cycles.
Regards,
Ian
Ian, no link to dataset, no answer. Sorry.
w.
Took seconds to find this:
http://wattsupwiththat.com/2012/06/25/an-interesting-graph-showing-solar-cycle-el-nino-and-surface-temperature-correlation-in-australia/
Milodon, I’m sorry, but I asked for a link to a dataset, not to a WUWT post.
w.
And Ian, my objections at the time still hold. The data you point to shows no sign of being connected to sunspot cycles.
w.
Willis,
Have you studied the Uppsala and Stockholm temperature series provided by the Swedish Meterological and Hydrological Institute?
These datasets show a decadal-scale cyclical variation, and have the additional benefit of being two of the worlds longest accurate temperature measurement series starting in the 1700:s.
http://www.smhi.se/klimatdata/meteorologi/temperatur/uppsalas-temperaturserie-1.2855
http://www.smhi.se/klimatdata/meteorologi/temperatur/stockholms-temperaturserie-1.2847
Thanks, Johan, I do love good datasets. I had looked at the Stockholm but not the Uppsala temperatures. Here’s the periodogram of the Stockholm dataset:
As you can see, there in no cycle from 5 to 50 years that exceeds a few percent of the range of the data. I don’t see the “decadal-scale cyclical variation” you mention.
My best to you, thanks for the data,
w.
If you ever want to be taken seriously by very serious climate scientists you should ditch looking at datasets and start running computer models. I’m sure you can come up with a computer model that spits out something resembling an 11-year cycle, which then becomes proof that your hypothesis must be correct.
It doesn’t have to hindcast the past, either. If you suddenly find a brand new computer generated 11-year cycle where one has never existed in any similar historical datasets, you’ve stumbled onto solid proof that manmade CO2 has made the earth more susceptible to sunspots, causing “more extreme weather”.
Love the spoofing of the so called climate scientists and their methods. I’m glad Willis is honest. It would be very helpful to know what has caused the global temps to stop rising, but no one really knows. What we do know is that CO2 isn’t the driver the so called climate scientists thought it was, and they should start looking elsewhere. But that doesn’t scare people into funding them.
I have a data set with some 11 periods over a 120 years. Not perfect spaced with 11 years interval .
It is annual see-level rise from Sweden- SMHI. You don´t get it if You look at the numbers on a yearly base. You have to treat the numbers gently, by a five Year equalization method . More on that on the site.
Hope Your Swedish is good!
http://www.klimatupplysningen.se/2013/10/13/angaende-historiska-havsnivaer-vad-doljer-sig-i-siffrorna/
Lasse, I looked at the sea level rise in Sweden (among other places) here. However, I didn’t display the Stockholm results at the time, so I’ve done them up anew:
As you can see, there is a 1-year and a 6-month cycle. However, there is nothing solar-related that is visible.
w.
Thanks!
It is a long way between the sun and the Baltic see-level. The currents draw energy from the Pacific that is the main reserve of energy. But anyhow I got a nice periodic pattern with now acceleration in recent years.
Willis,
A simple question. Given that “The climate system is a coupled non-linear chaotic system” (TM IPCC)
Would you expect inputs to a coupled non-linear chaotic system to provoke the same response each time they were applied? This is as foolish as creating linear projections based on the past behavior of a coupled non-linear chaotic system.
What one would expect is that each periodic input generates some effect but that the exact effect would vary dependent on the orbits around the multiple interactive attractors and the state of other coupled inputs and effects. In other words the response of the coupled non-linear chaotic system will be different each time. That does not mean that there is no effect, there will always be an effect from (response to) an input, but it does mean that looking for simplistic responses to that input in one or more metrics is a nugatory effort.
An effect that is so small it cannot be detected in a noisy signal can thus be essentially set at 0 then? For example, when the equatorial trade wind is blowing from East to West, we get a nice walker cell circulation and a very predictable cloud formation in Indonesia. But when it calms down or even goes the other way in the form of Westerly Wind Bursts, we get another kind of less organized and less cloudy cloud pattern. It’s a big effect and models match observations. One can input solar metrics into the mix or not. It doesn’t change the outcome. Therefore, no matter what is happening with solar input, that part of the equation is so miniscule that one can ignore it. In chaotic systems that are being transformed into calculations, models, etc, one can do that mathematically and be safe in doing so. It’s the kiss principle.
The correlation between increased CO2 and increased water vapor is one of those types. It has not been observed to match the model. Why? Because when switching out CO2 for things like ENSO, the correlation with water vapor change is strengthened to a much greater degree than when using just the CO2 data. This particular well-known and stronger correlation is not mentioned in paper after paper proposing that we are all going to die from CO2.
>> An effect that is so small it cannot be detected in a noisy signal can thus be essentially set at 0 then?
Pamela, this is incorrect. There are plenty of systems that are complex enough that variations in input are not found in the output. This doesn’t mean that the input can be ignored or doesn’t matter.
I already gave you the great lakes / rain example. There are many examples of this. People eat in 3 discrete pulses per day, but we shouldn’t expect to find evidence of those pulses in weight measurements. However, it’s obvious that increasing the amplitude of those pulses will add weight.
Other examples include anything that uses pulse-width modulation. Electric cookers use a thermal oscillator running at approximately two cycles per minute, while the duty cycle varies according to the knob setting. The thermal time constant of the heating elements is several minutes, so that the input signal variations cannot be found in the output temperature.
Yet, even though the same attribute is present for this very simple thermal system, some commenters here seem to consider the lack of output variation in the much more complex thermodynamic system of the Earth (core/crust/land/sea/atmosphere) as evidence of something.
Viking, your eat more weigh more simile bears no relationship to solar-temperature cause and effect. The percent change in solar output at the Earth’s surface is nothing like the power of a single calorie on a human body. However I will concede that oceans have the unique capacity to hold onto heat hence build up heat, much like a sedentary lifestyle can hold onto and build fat from calories that are not burned off.
Pamela, I wasn’t implying that the cause-effect mechanism was the same, although both are definite and undeniable periodic injection of Joules. My analogy was that both systems are complex enough that one should not expect input variation to be visible in the output. Wouldn’t you agree that one would be laughed at if one claimed that food intake was irrelevant to obesity because there was no meal signature in weight data?
VikingExplorer November 1, 2014 at 11:20 am:
+++++
What you have been writing makes good sense. I appreciate your input here. Thank you!
Right you are, but .. how do you propose to detect such a quasi-random response?
It is necessary to understand the systems and how they interact and in a coupled non-linear chaotic system of coupled non-linear chaotic subsystems when many of the parts of the system and their relationships are unknown, or misunderstood and their effects less so usually measured in the wrong units. So you can not detect that quasi-random response because you don’t even know what response you need to look for or when/where you need to look for it.
Ian W November 1, 2014 at 3:30 am
Hey, I’m not the one making the claims that the sunspot cycle affects the weather. I’m just the poor fool that is trying to find the evidence.
w.
“It is necessary to understand the systems and how they interact..”. Exactly. Let me quote from Rex Stout’s The Mother Hunt: There must be traces – letters or telegrams, check stubs and canceled checks, a hair from your head in her car, a hair from her head in your car – the possibilities are innumerable, now that you have been named.
Using statistics blindly won’t necessarily result in an insight. Once you have an insight, statistics becomes a powerful tool to support – or to disprove – it.
“…to seek, to find, and not to yield…”
‘ if the correlations are as great for sunspots lagging the climate as they are for sunspots leading the climate, you’ve got problems.’
If the sunspots lag the climate by eg six months, isn’t that the same as saying the climate lags the sunspots by ten years and six months, presuming that the peaks and troughs of the solar cycle are similar each time?
It is what I call a ‘sometime’ theory
http://www.vukcevic.talktalk.net/SSN-UKsh.gif
However I’ve been working on something I call Geo-Solar cycle (solar + earth’s magnetic field)
http://www.vukcevic.talktalk.net/GSC-UKsh.gif
It does indicate that might be something to it, but delay of 13 years is very hard if not next to impossible to explain, possibly to do with the two N. Atlantic gyres, since UK gets its clouds from there.
Just a thought.
.
GeoSolar cycle does a good job of reproducing the de-trended NA SST with an extra couple of years delay
http://www.vukcevic.talktalk.net/GSC.gif
However, delay is the proverbial ‘elephant in the room’
>> but delay of 13 years is very hard if not next to impossible to explain
Seems easy to explain, given that ocean currents in 3 dimensions are involved.
http://svs.gsfc.nasa.gov/cgi-bin/details.cgi?aid=3827
as an aside, it would be interesting to see that graphed against the north atlantic oscillation index
reply to your comment went in wrong place, see above above
thank you for taking the time to post that,interesting result . a reasonably constant delay could imply there is constant worth looking for ?
KTM’s “contribution” at 3.17am is one of those that annoy me when I am looking for meaningful contributions to the debate. Why bother “contaminating” the thread with such nonsense? I wonder why he/she even did not add “sarc off”? Did KTM actually mean what he/she wrote?
It is sarcasm, John 😉
Hi Willis. Thanks for another well-presented argument.
BTW, the Don Quijote armor looks good on you.
Cheers.
Forget about the solar cycle.
The gold mine here is the cloud cover dataset that might actually be worth using. Clouds are supposed to be going down as it gets warmer. One of the great questions remaining in the debate is the feedback response of clouds.
Nobody really believed the ISCCP cloud dataset before (even climate scientists) but maybe matching up the two different datasets so closely provides some backing that both are reliable.
>> Clouds are supposed to be going down as it gets warmer.
Wouldn’t warmer temperatures result in more evaporation, which leads to more clouds?
Nope, not in this theory.
Low cloud cover goes down, high cloud cover goes up but the net impact is a reduction in the degree of cloudiness producing a positive feedback as it gets warmer (less solar reflection). There is more water vapor in the atmosphere but less net cloud cover somehow.
[This implies that during the ice ages or during snowball Earth, for example, there was a huge increase in cloud cover which, as you note, makes no sense whatsoever].
Without the positive feedback from clouds (let’s say it is actually net zero), the total temperature increase from global warming falls from 3.0C per doubling to 2.0C per doubling. Hence, they need to keep it in the positive illogical range to keep the disaster scenarios going.
I’ve been meaning to do this chart for awhile. A depiction of how the feedbacks (and their assumptions) are crucial to the global warming theory. Change the feedback assumptions by modest amounts and the whole amplification process falls apart.
http://s29.postimg.org/5x7dto207/Global_Warming_and_Feedbacks.png
Bill, different kinds of clouds have different effects. For example, high thin cirrus has low space oriented albedo to income solar but high earth oriented OLR albedo because comprised of ice rather than water. Total cloudiness per se would not be expected to yield much. And Willis composite suggests that indeed it doesn’t, at last for the geography charted, which is basically CONUS. Probably too regional to detect any global climate signals, most of which are likely ocean oriented, as that is where Earths heat sink and water vapor feedback sources originate.
Bill Illis November 1, 2014 at 4:47 am
I couldn’t agree more. The models say that cloud feedback is positive. My hypothesis that emergent phenomena control the temperature says that cloud feedback is negative.
This dataset strongly supports my hypothesis.
w.
Willis said:
“I draw no great conclusions from this, except that overall, as the US warmed post-1930, the US cloud coverage rose as well. Go figure …”
The US cooled from the 1930’s. Cloud then cover reduced from the 1970’s as the US warmed again:
http://wattsupwiththat.files.wordpress.com/2011/10/palmer_figure2.png
And cloud cover increased again in the last decade as US temperatures have fallen:
The observed trend in cloud cover is the most interesting result. The fact that the ISCCP data agrees so well with the ground observations is very striking.
Agreed, I was quite surprised by how well the ground data matched the satellites.
w.
I have often postulated that the clouds around the equatorial belt may have far more predictive power on ocean heat than global cloud data. This is because of the potential depth to which shortwave infrared is absorbed when the angle is perpendicular versus glancing. It would be of interest to compare equatorial cloud data versus upper oceanic heat measures all centered solely around this equatorial belt. The more poleward the data, the less it is useful in making predictions in terms of future oceanic heat increase.
I speculate that the amount of SWIR, and the depth of it, that bore upon the equatorial belt during the last substantial La Nina is to blame for current North Pacific SST increase. Therefore, equatorial La Nina and possibly long lived La Nadas could be useful metrics in forecast future SSTs. Combined with oceanic currents, it is possible to even forecast where and when these equatorial sourced increases in sea surface temperature will surface.
Wouldn’t albedo in the equatorial regions be the most significant measure?
Robert of Ottawa November 1, 2014 at 5:20 am
Seems like it to me. I plan to look at the ISCCP dataset vis-a-vis the CERES dataset, but it’ll take a bit of programming.
w.
Willis, svensmarks cloud cover would be upper troposphere cirrus and alto stratus – please cleanse your data set of all low altitude cloud types and see what you get
James of the West November 1, 2014 at 5:27 am Edit
I would if I could, James, but the problem is that the ISCCP dataset only goes from 1983 to 2009, which is only a couple of sunspot cycles. As a result, we can’t conclude much about the effect of GCRs, there’s simply not enough data.
w.
Willis, the mystery isn’t that there is a 11 year signal in the climate there MUST be, the mystery is why there is not! There should be, and based on summer winter cycles it should be about half a degree, but uniquely for sunspots its not. It works for earth and solar system geometry but seemingly not for Emission. Why not! That is the question to be solved. Non linear feedbacks are a possible answer but why doesn’t that cancel geometric forcing changes?
The answer may lay in other data, there are saturation effects in climate, perhaps you need to look beyond (min + max)/2 which is about as dumb a view as you could get, perhaps the sunspot cycle shows up in the rate of warming at dawn or dusk or relationships between summer and winter. Perhaps there is an ionospheric or stratospheric effect that counters warming due to the increased insolation, eg such as say increased jet stream velocity. Maybe the increased insolation reduces cloud cover at night which allows more radiative loss. Could be anything. Clearly though (Min + max)/2 isn’t going to solve that puzzle.
One hint though, geometric forcings changes the distribution of sunlight over the earth, sunspots do not – this difference might account for somthing. Non linear feedbacks might affect sunspot warming differently to that caused by geometric changes.
Bobl
Imagine kicking someone’s ankle once every 11 seconds – almost every time you do it there is no effect but by chance if you give the ankle a light tap just before the foot hits the ground they fall over then you carry on fruitless kicking out of a few hundred attempts one fall. So you would say all those times the ankle was kicked no reaction one random time something happened – there’s no such thing as an ankle throw. You would be mathematically/statistically correct but factually wrong. You just do not understand the system that you are measuring
I don’t see your point here, if you increase insolation you must increase temperature or the earth would be a frozen mass at 0 K, this is known to be true. The question then remains why can’t this signal be extracted over 20 cycles (even with niose reduction of some 20 times). The obvious answer is that there is a counter forcing that applies to insolation but not geometric forcings, that could he a feedback or a non-linearity ( forcing is too weak to shift the climate to a new attractor), but it could also be a related effect, say solar wind changes, UV or magnetism resulting in some atmospheric effect.
I agree with you in the sense that the sunspot cycle mystery shows we don’t understand the system we are measuring.
By the way the issue you point out pervades medicine, results are based not on biochemical response to drugs but rather the statistics of a double blind trial. Confounding factors (say side effects) can and do affect outcomes, efficacious treatments are being elliminated all the time by the mechanism you mention, the few percent it helps are hidden by the morbidity of the side effects. It also turns up in cancer nomograms. The doctor will cite a nomogram that says half the people who contract your cancer will be dead in five years! That sounds really bad, except when you realise that this figure is all cause mortality and that older people (80 and up) are overrepresented in the population, most of them actually die from something else! The wrong statistic is being measured, what you really need to know is disease specific mortality, not all cause mortality.
“you can’t prove a negative in any case”
BS. Evidence of absence proves a negative.
From Wikipedia:
If someone were to assert that there is an elephant on the quad, then the failure to observe an elephant there would be good reason to think that there is no elephant there.
By the same token: If someone were to assert that there is an temperature signal in the clouds based on sunspot cycles, then the failure to observe a temperature signal there would be good reason to think that there is no temperature signal there.
Always be careful when quoting Wikipedia. The elephant in the quad has no boundaries set on it. You may be able to eliminate large elephants, but tiny toy elephants in the bushes might be missed. It depends on how thorough you are. It also depends on what the application you’re studying elephants for. If you are in an area where elephants naturally live, its helpful to know where they are to avoid being trampled by one.. One observation won’t prove they aren’t there all the time, only the time the observation was made.
The so called climate scientists may be trying to associate the current CO2 rise with current temps. But if CO2 is rising from both man made causes and natural causes, there may be a lag of a few hundred years. If the oceans are responsible, the cause of the current warming may have ended a long time ago, and Trenberth’s missing heat may have long ago been emitted to outer space, and the cycle is about to turn cold. They may be searching for a current cause when there isn’t one because they are looking in the wrong place.
Willis writes:
“My plan for the quest has been simple. It is based on the fact that all of the phenomena commonly credited with affecting the temperature, such as cosmic rays, the solar wind, changes in heliomagnetism, changes in extreme ultraviolet (EUV), or changes in total solar irradiation (TSI), all vary in phase with the sunspots.”
If only it were that simple. The solar wind for the last 50 years did not follow the sunspot cycles very well at all:
http://snag.gy/vM2dv.jpg
The previous 6 cycles have a fairly regular low value around a year after each sunspot minimum, but there is often a low in the solar wind speed at sunspot maximum too, which is strictly out of phase with sunspots, and radically alters the periodicity of lows in the solar wind speed.
http://www.leif.org/research/Ap-Monthly-Averages-1844-Now.png
Cloud cover in the US will be effected by the AMO mode, which appears to switch phase in relation to sunspot cycles depending whether it is in its cold or warm phase. That would tend to cancel out a solar cycle length cloud response signal in data sets longer than 3 solar cycles:
http://www.woodfortrees.org/plot/esrl-amo/mean:25/plot/sidc-ssn/from:1850/normalise
There is also the problem that warmer in summer can be less cloudy, but warmer in winter can be more cloudy. And what about the regional differences in change in cloud across the US, by averaging many stations you may be losing the signal, as some regions may get cloudier while cloud reduces in other regions.
Ulric, what is it with you and links? Without a link to your solar wind data, there’s no way to determine anything about your claims.
In any case, you claim:
In fact, a periodogram of the OMNI solar wind data clearly shows the sunspot cycles:
Solar Wind Data Source: ftp://spdf.gsfc.nasa.gov/pub/data/omni/low_res_omni
As you can see, the strongest cycle in the solar wind (red), twenty percent of the range of the variations, is at around ten years. This of course is clearly related to the solar cycle (blue) for the same period, as you can see by the correspondence between the two periodograms.
In other words, your claim that the solar wind doesn’t follow the sunspot cycles is simply not true.
w.
“Without a link to your solar wind data, there’s no way to determine anything about your claims.”
The graph is there in my comment, a visual analysis immediately shows a poor match. And your periodicity result of only 20% confirms that the solar wind speed does not follow the sunspot cycle well during the last 50 years.
“In other words, your claim that the solar wind doesn’t follow the sunspot cycles is simply not true.”
With the solar wind speed in sunspot cycles 19 & 20 being lower at sunspot maximum than at the following sunspot minima, it is decidedly true.
Ulric: The problem here is that the solar wind data does have a strong solar cycle dependence, but that the shape of the cycle is different from that of the sunspots:
http://www.leif.org/research/Solar-Wind-Climatology.png
The curves show the solar cycle behavior through the cycle based cycles 13 through 23. The average cycle is shown at the left, and then repeated five more times to the right to make the cyclic behavior more clear.
So, if you look for periods, you’ll find the 11-yr cycle.
Leif, as you know there are large drops in the solar wind at the sunspot maximum of some cycles, e.g. SC’s 20 & 21, that radically alters the periodicity of major drops in solar wind speed through the last 50 years. Can we pay attention to what has actually happened at these events rather than meaningless averages thanks.
Ulric: Can we pay attention to what has actually happened at these events rather than meaningless averages thanks
Perhaps you should do that. The average behavior during the space age is very closely the same as for last ten solar cycles. Averages are not meaningless, they dampen out the noise and let the truer signal through showing what is really repeatable and stable. You can see that from this comparison of the spacecraft area and the full dataset since cycle #13:
http://www.leif.org/research/Averages-Space-Climate.png
You could redeem yourself by acknowledging here that you agree and understand his.
“Averages are not meaningless, they dampen out the noise and let the truer signal through showing what is really repeatable and stable.”
They dampen out the noise in your graph, the real world responds to all the noise.
“You could redeem yourself by acknowledging here that you agree and understand his.”
I already knew about the averages, you could redeem yourself by acknowledging the noted lack of coherency of the solar wind with the solar cycle through the last 50 years.
noted lack of coherency of the solar wind with the solar cycle through the last 50 years.
The Ap-index [you showed it yourself] is a good measure of solar wind speed. The last 50 years show very clear solar cycle period dependence with a peak every ~11 years during the declining phase of the cycle [as is also shown in the average curve]:
http://www.leif.org/research/Ap-index-peaks.png
So, you will note that there is no lack of ‘coherency’ in the data. Perhaps in your persistent claim instead.
The solar wind speed the last 100 [and 50] years show a very strong ‘coherence’ with the solar cycle always maximizing during the declining phase of the cycle with peaks spaced the cycle-length apart:
http://www.leif.org/research/Solar-Wind-Speed-Last-100-Years.png
This should put the matter to rest, although I doubt you will admit it.
I am well aware of the peak in solar wind speed on the declining phase of the sunspot cycle, the lack of coherency that I had noted above was the frequency of slow SW periods in the last 50 years.
The last 50 years are not any different than the rest.There are 23 years with an average [404 km/s] less than the average speed [430 km/s] compared to 18 years before the space age with an average [401 km/s] less than overall average. Hardly a significant difference. Said in a different way: the average speed the first 50 years was 444 km/s vs. the average of 441 km/s for the last 50 years. The last 50 years are not any different from the first 50 years over the 100-yr period. And both show a strong ’11-yr’ cycle with wind speed peaking during the declining phase in all cycles.
lsvalgaard
November 3, 2014 at 1:52 pm
“You are more than usually dim today.”
Given how long it has taken you to address the low SW points that I am discussing, I can only regard your ad hominem as a projection.
“If anything, the recent period has fewer low-speed Ap periods.”
There were unusually deep SW lows for sunspot maxima at 1969 and 1979/80, but not at 1989 and 2000, trashing any regularity in the last 50 years:
http://snag.gy/hSqT4.jpg
There are lows and highs of short duration all the time for someone [like you] to cherry pick what he likes. The fact is that the general character of the variation follows the solar cycle very well with a peak on the declining branch in almost every cycle. The last 50 years are no different. You have painted yourself into a corner and can’t afford to creep out. This is clear to everybody so no more demonstration is needed.
No, the point is that short dips in any quantity occur rather at random and thus do not influence the fundamental solar-cycle variation of solar wind parameters. The sooner you realize this, the sooner you are to enlightenment
“The last 50 years are not any different than the rest.”
Again, for the frequency of the major periods of slower solar wind speed, they do differ, as can be readily seen on the Ap index series above.
You are more than usually dim today. For you convenience I re-show the Ap-index so you can see that there is no real difference. If anything, the recent period has fewer low-speed Ap periods.
http://www.leif.org/research/Ap-index-peaks.png
Note the low speed around 1900, 1913, 1924
lsvalgaard said
“You are more than usually dim today.”
Given how long it has taken you to address the low SW points that I am discussing, I can only regard your ad hominem as a projection.
“If anything, the recent period has fewer low-speed Ap periods.”
There were unusually deep SW lows for sunspot maxima at 1969 and 1979/80, but not at 1989 and 2000, trashing any regularity in the last 50 years:
There were unusually deep SW lows for sunspot maxima at 1969 and 1979/80, but not at 1989 and 2000, trashing any regularity in the last 50 years:
Not at all. Note the red arrows spaced a solar cycle apart:
http://www.leif.org/research/Ap-index-peaks.png
or will you persist being dim?
There is nothing unusual about 1969 [left red oval]:
http://www.leif.org/research/Solar-Wind-Ulrics-Folly.png
There were also several such dips around the maximum of 1990 [right red oval] and many more besides. The solar cycle variation of the solar wind speed is clearly shown by the blue ovals spaced a cycle apart. Willful dimming is not helping you much. Hope this final demonstration makes closure possible for you.
Your red oval encompasses 8 years, and even includes the low in the solar wind near the start of SC 21. The point is that the lows tend to happen at roughly alternate 3.5 and 6.5 intervals, in and out of phase with the solar cycle.
Thanks Willis. I like data, rather than simulations. As a complete novice, could you answer a couple of points.
If the cloud cover varied in cycles from 9 years to 13 years (ie: +\- two years), would your analysis detect it?
If the controlling factor was Sunspot cycle length, rather than Sunspot amplitude, would your analysis detect it? (Cycle length would be multi-decadal, rather than the decadal amplitude cycle.)
Thanks,
Ralph
ralfellis November 1, 2014 at 6:29 am
Yes. Look at Figures 1 and 2. The cycles vary, but the periodogram shows them nonetheless.
I think so, but If you have a dataset where the controlling factor is the length of the cycles in forcing we can test it. Me, I don’t understand the physics of how such a cycle-length based control would work in the climate system, so any information about that would be helpful as well.
w.
>> I don’t understand the physics of how such a cycle-length based control would work in the climate system
w, how could you not understand the concept of pulse width modulation? All of our homes are heated this way.
Why don’t you take a break from searching for evidence of solar variation and try a system a million times less complex. Measure the temperature in your house periodically and look for evidence of the input pulse frequency. If you can’t find it, then I’m wrong about how our homes are heated.
Wait…
er…
I thought that logic was so good…
Electronics math works for real systems’ frequency responses. I think engineers tend to see the climate as very complex inductors, capacitors and resistors, hence RLC circuits. PWM uses a given frequency of digital pulses, where the percentage of on-time is varied. Light dimmers are done this way… That allows you to excite the system with the correct voltage but limits the overall power from entering the system. Using a variable resistor would limit the the voltage, which can cause problems for systems designed to run at a certain voltage.
I understand very well how pulse width modulation works in electronics circuits and in warming the house. What I don’t understand is how it would work in climate, given the evidence we have.
That is to say, we can find no evidence of the sunspot related cycles in the surface weather records using e.g. a cross-correlation analysis. Now, this analysis specifically includes the variable cycle lengths … so IF the variable cycle lengths are regulating the temperature/sea level/rainfall/whatever, the CCF analysis should show it.
But it doesn’t … so again I ask, how is this supposed to work in the climate? As Viking says:
That’s great. So let me answer back.
w.
VikingExplorer November 2, 2014 at 10:04 am
Why don’t you take a break from searching for evidence of solar variation and try a system a million times less complex. Measure the temperature in your house periodically and look for evidence of the input pulse frequency. If you can’t find it, then I’m wrong about how our homes are heated.
Now, there have been several comments in this thread that purport to ‘prove’ that there is a solar cycle variation in any number of parameters. So, by your ‘logic’ all those peer-reviewed papers are just plain wrong. It would be good for you to state that explicitly here and now rather than just sit silently and less all those correlations pass you by.
Bob Weber said:
“Today’s solar flux dropped to 121. My analysis of SST and solar flux from 1960-now indicates the oceans on average drop in temperature when flux reaches 120 sfu/day, and the temps increase above that level of solar flux.”
That would fit with my hypothesis that a higher solar flux reduces ozone above the poles in the stratosphere. That cools the stratosphere above the poles which causes tropopause height to rise and the climate zones and jets to shift poleward.
The result is lower global cloudiness and more solar energy into the oceans.
The reverse when the solar flux is low.
I would be very intrigued if Bob or anyone else could show a rapid short term response since I have been focusing on much longer timescales.
It would be very useful if Bob has identified the switchover point at 120sfu/day
Stephen thanks for your vote of confidence. The rapid short term response is most readily observable in a single solar rotation when one half of the rotation has little to no activity, followed or preceded by a much higher activity level, such as what happened over the past 27 days, and will most likeky happen again in the next 27 day rotation.
It’s doubtful anyone else has made this connection as I’ve not heard anyone here or elsewhere talk about this even after I’ve mentioned it here several times over the course of the year.
My answer here also applies to Bruce Ryan’s comment further down the line here.
Solar influences the weather and climate two ways: variable radiative output in sunlight; and the particle effects from solar plasma outflows that operate electrodynamically, what I call “electric weather effects”.
Stephen I’ve been building a case all year long for what I’m saying, and working towards first publishing blog post articles here and elsewhere if our host will post them, and hopefully later science paper(s) for submission to a journal for peer-review. There’s so much more to say but I want to save it for those articles and papers for now.
Since I value the input from people across the spectrum I look forward to getting your feedback and from others when that day arrives.
The Sun causes warming, cooling, and extreme weather events, not CO2.
Good luck to you (and me) 🙂
And so much for the unnecessary and ostentatious comment about “hand wringing”.
Bob, you haven’t put forward a single link to allow us to investigate your claims, and despite that, already you’ve gotten a “vote of confidence” … must be nice to get people to believe you without a scrap of evidence.
Me, I’m not that foolish …
w.
Willis I’ve been out cutting wood all day and I just got home. Chill.
Except Stephen, you have cause and effect reversed. It is ENSO conditions that shifts the polar jets. We now know that ENSO conditions have both short and long term oscillations. You need to complete your study by determining whether or not there is a lag between what you think is the cause and what you think is the effect. I can tell you upfront that researchers have already established that ENSO changes first, which THEN shifts the polar jets. You must therefore focus your speculations on causes at the equator, not the poles.
http://www.srh.noaa.gov/srh/jetstream/tropics/enso_impacts.htm
There are clouds, and there are clouds. High level ones interact with temps differently than mid and low level ones. So this throws three more variables in the mix.
Does increased solar activity result in more t-storms giving more high level clouds? Or does it cause more or fewer low level clouds to form in the mid and upper latitudes?
There really are so many variables in all this. it seems a little fruitless to try to find patterns in just one or two.
Wouldn’t Fourier Analysis reveal major cycles?
The periodograms I show above are done using Fourier analysis, which will indeed reveal major cycles.
w.
I wasn’t sure how you constructed them, but they looked quite laborius. So i tried a shortcut version of your analysis with a single station with, was stunned that it really showed the 10.9 year cycle on the first try and was planning more sober tests on the other station later.
Now i realized that it was not reproducible even for the same station with other years than 1922-1987.
So i don’t even try the other stations anymore, especially since you already did it.
Thanks
I would think one would find solar influence in the ocean temperatures over time rather than land if it existed.
Ozone breakdown, stratospheric cooling and mixing atmospheric layers and influencing jet streams and ultimately tradewinds etc. Unfortunately, no data set that I can find.
forgive me the concept, would the relative position of the earth in its orbit have influence on the strength of whatever the sun is doing? iow would you need to have a chart for every month of the year, or a chart for each of the summer months of each hemisphere?
Trouble is the position of the sun on the earth is so short lived at any orbital position the fluctuations in the suns spots would have to be perfectly nil in order to witness a correlation.
This is all academic. I can’t see why people can’t accept that the climate is a very chaotic system, and does not really respond to any of our little pet theories. As such, you’ll get this beautiful correlation and model results … “sometimes”, and at others, complete divergence. Hold the data just right … shake it appropriately, and use the “appropriate” mathmatical analysis, and something will always fall out …. or not.
CO2 is not the culprit
Solar is not the culprit
Aliens are not the culprit.
Math can’t explain it …. nor can “People Magazine”, or Time or any other journalistic endeavor.
BUT .. .alas .. our little egos just won’t let it go.
Yes and no. We are making small steps toward understanding atmospheric/oceanic teleconnections at the local and regional level. Of even more importance we are making small steps toward understanding the bridges between one area of the globe and another area of the globe (IE ENSO 3.4 and say…the Atlantic Multidecadal Oscillation, or upward propagating Rossby Waves and Jet Stream behavior). What models get wrong and Earth seems to care little about is long term forecasting 30 to 50 years out. All such forecasting endeavors that make a difference to us at local and regional levels get washed out in the error bars. Most are wrong and the ones that are right got there by sheer chance.
The IPCC does consider warming as caused by aliens. Since they separate human activity from what they refer to as “natural” we become unnatural, and thus alien.
Willis, I was hoping you would show proof of the absence the sun spot period in the cloud data. Instead of highlighting the the location of the required 11 and 10 year periods, and showing their absence, you leave the reader to struggle with the peak locations of your oddly labeled graph. And arguing that “nothing in the range from five to twenty-five years exceeds about 5% of the total signal” is not equivalent to proving that there is nothing there. And when was the last time that anybody here accepted the argument that trends less than 5% of the seasonal variation can be dismissed as irrelevant. Indeed, the alarmists make the opposite argument.
On the other hand, if there is something there, that would also not be proof of causation. I would also like to see a comparison of your averaged cloud data with averaged temperature data. My hunch is that there is only superficial similarity. I don’t see the 70’s cold spell nor the recent plateau or pause in your cloud data.
I can see that as a loop. If A then B, if B then A. In programming this is called a run-away loop and can freeze a computer. Plugging in clouds and cooling for the variables, is it reasonable to think clouds cause cooling and cooling causes clouds? Probaly not, it is more likely clouds causes cooling and cooling causes some number of variables in a some squence to cause clouds.
In PC operating systems the idea of linear processing went out the window a while ago, operating systems are event driven. Nature is similar to this in that it is one big loop constantly reacting to and creating events. Events occur concurrently, can be cascading, are not predictable (the operating system does not know when user A might start program B), and can clash. Within the big loop there are smaller nested loops doing the same thing, shooting events outwards eventually reaching the outer loop.
Causation is relatively easy, If A then B. Understanding causation has shot civilization forward. However, in terms of understanding how climate works, causation is like learning the alphabet, the first step but still far away from making meaningful stories. Earth and it’s atmosphere is a complex, multi-layered, event driven system with randomness and unpredictability a core component. At this point working on climate models helps understand the events and event handlers, claiming models simulate Earth and it’s atmosphere is science-fiction, claiming to be able to predict climate for centuries into the future is both childish and a God complex.
@Willis
Your hypothesis that “tropical clouds and thunderstorms actively regulate the temperature of the earth” suggests you should look at the rate of cloud cycling vs.sunspot number. What you are currently doing is looking at total cloud cover. I suspect all the interesting stuff has been averaged out of the dataset. Why not try again eating your own dogfood, as they say? 🙂
Who on earth says “eating your own dogfood”, and what does it mean?
w.
It specifically means “using the software one sells to other people”, or more generally, doing as you tell other people to do.
Thanks, Viking, but I’ll wait for JDN to answer my question about what he means by it. Your interpretation doesn’t make a whole lot of sense … what have I been “tell[ing] other people to do”?
w.
Wow. I didn’t know that was going to be controversial.
The phrase “eating your own dogfood” is usually used by programmers to imply that they get their work done using the tools that they themselves made. It’s stretching the metaphor a bit, but the tool you made is an insight that storms seem to regulate temperature. Since that’s your thing, I though you might look for correlation of the global thunderhead area with sunspot number, or something to that effect.
I’m just as puzzled as you are about why there is no periodicity.
Thanks, JDN. I’d just never heard that “dogfood” expression.
I’d love to but as far as I know, there is no measure of global thunderhead area, although I could be wrong. I’ll have to look deeper into the ISCCP data, since that information may be there … so many datasets, so little time.
All the best, thanks for the suggestion, and for clearing up the confusion.
w.
Would a global lightning database serve as a proxy for thunderstorm severity?
See http://thunder.nsstc.nasa.gov/data/data_lis-otd-climatology.html
The cloud cover data that you have ‘uncovered’ for the US seems much more interesting. If one looks at this from what might be called a ‘gaia’ perspective, it’s mightily logical that cloud cover would be the thermostat. I can’t believe that nobody has taken a good global look at this, but even local data ought to yield some very interesting results without too much torture. There is likely good, long-term cloud reporting in the UK to link with the very long CET temperature record, although as the UK is subject to very rapidly changeable weather patterns on the edge of the continent this may limit what we can learn.
Part of the problem is using the Average SFU to represent the Solar Output energizing the Earth. One needs to use the “area under the curve” or the integral. This gives one number that more accurately represents the total energy for that 10~11 year period.
To be more precise than that, one needs to take into account the oceans and their heat storage and release.
To that end, I made a model, based on physics, to merge the Solar UV and Ocean Temperatures. I ran this model from 1650 until now.
Simply stated:
1) EUV > 120 warming.
2) EUV = 100 to 120 holding steady
3) EUV < 100 cooling.
Note that the EUV values must be averaged [area under the curve] over the 10~11 year solar cycle.
Now, if the Sun is in a reduced output period, then there will be less EUV resulting in less Ozone. This will allow the Ozone holes to expand. Since Ozone blocks heat flow into space [absorbs infrared and re-radiates 1/2 to earth], the holes will allow heat to escape at an increased rate. The effect could be similar to falling off a cliff, Global Temperature wise.
We are now into "only" 15 years of reduced Solar Output [starting at 2000]. Look at the results in the Northern Pacific and Northern Atlantic:
http://weather.unisys.com/surface/sst_anom.gif
Also, one can see a signal of the EUV in
http://ocean.dmi.dk/arctic/meant80n.uk.php
Notice the temperature "peak" at day 300. This was due to the huge EUV output from the giant Sunspot.
Also, one can track the Ozone holes via:
http://ozonewatch.gsfc.nasa.gov/monthly/NH.html
You have identified what I have identified. Congratulations! In fact, I also use Unisys!
jlurtz November 1, 2014 at 7:26 am
Part of the problem is lack of links … exactly what are you calling “Average SFU”, and where is the link to the “SFU” dataset you are using?
Sorry, but that is reanalysis data. In any case, what is the dataset which you are using to determine the variation in EUV during that period? I ask because as far as I know, the EUV didn’t hit the earth until October 24th … but the reanalysis data you show has the temperature starting to rise on October 20th. So I’m curious about how a temperature rise on the 20th can be caused by a rise in EUV on the 24th.
w.
w.
There are no data sets. I make and analyze my own. I use
http://www.solen.info/solar/
as my inputs. I did this because the centroid of a triangle is 1/3 the way from the base, but the average is 1/2 the way from the base! The amount of energy is strongly related to the centroid not the “instantaneous average”.
As per the rise in temperature related to EUV, the EUV is radiation and it absorbed by the ozone layer and remitted [we measure the remitted as SFU]. The amount of radiation absorbed is a function of the path length: short path length less absorbed; long path length more absorbed. The Arctic this time of year has a long path length.
So, the EUV does not need the peak to affect the ozone, it is more a function of the “path length time” that the ozone is exposed to the EUV.
What is wrong with “reanalysis of data”? All data is analyzed and then reanalyzed, etc.!
Progress is made by new ideas and better analysis; not in rehashing the existing muck.
You will get ‘updated’ UNISYS interpretation, note that I am NOT saying it is correct, by using the proper link. It is now SSt anom_new as opposed to SST anom, which you show above.
http://weather.unisys.com/surface/sfc_daily.php?plot=ssn&inv=0&t=cur
Sigh. No, we cannot. All that we can state is that there is no direct univariate correlation, and since correlation is not causality that is insufficient to justify the conclusion, just as it would be insufficient to draw the conclusion that they are “affecting” (the cause of) observed variations in the dataset if there were correlation.
. Bayes theorem comes into play.
is uninteresting and not predictive, but this does not suffice to show that 
is not causal until you show, or argue, that 
is independently uninteresting, that 
is uninteresting, etc.
th dimensional?
But in a highly multivariate system this isn’t just an idle repetition that can safely be ignored. It is serious stuff. The climate is “probably” orbiting in poincare cycles in some high dimensional space in association with a strange attractor. The temperature, cloudiness, etc etc are all projections of this orbit. The orbit itself is perhaps not some simple hyperplanar single loop in two effective dimensions, because that would produce precisely the kind of signal you are looking for the projective dimensions of all of the variables contributing to the hyperplane and you’re not finding it, but then, why would it be?
Even in the simplest 2D systems, predictive modelling with 2D outer-product logistic regression fails because sometimes the relationship between the 2 variables and the outcome is of the exclusive or type, not the independent logistic type. In statistics-speak, the relationship is not separable, the joint probability distribution does not factor,
All you are showing is that independently,
I think that you’ve already made assertions for causal relationships of precisely the form that would confound this yourself in the past. Water warming because of e.g. seasonal changes produces more water vapor (a potent greenhouse gas). But more water vapor makes more clouds (a potent high-albedo covering and latent heat transportation system). The greenhouse warming is offset by the albedo cooling and the climate is (likely to be) far less sensitive to increases in water vapor content caused by warming or cooling variations than one expects by considering only the direct relationship “more water vapor means more warming”, or the direct opposite relationship “more water vapor means more cooling”. This is further compounded by the fact that this same phenomena could (still) be net warming or net cooling systematically, but only in a way that depends on average latitude, or a way that depends on average latitude and whether one is over certain kinds of surface, or a way that depends on average latitude, whether one is over certain kinds of surface, and time of year/local mean surface temperature!
Seriously. Water vapor and consequent clouds and cooling might well dominate heavily over otherwise dry tropical or subtropical deserts, be irrelevant (on average) in the temperate zone, and be net warming in the polar regions, but only over ice where the increase in albedo is invisible. You might see a very clear signal if you look at these regions independently, but by the time you average the warming and cooling globally the signal all but vanishes. The systems dynamically redistributes many things, to be sure, but perhaps the average temperature, the average total cloudiness, and so on, do not change (much) as it does so.
This is why the climate system is such a bear. We have inadequate data to do a particularly good job of trying to untangle the joint distribution a mere two or three variables at a time, let alone the five or six visible in the example, and that data becomes more inadequate rapidly as one goes into the past, to the point where for more than thirty to fifty years ago, one isn’t even modelling the data, one is modelling a model that is supposed to reconstruct the data we do not have for that era subject to some assumptions, which then become additional unstated Bayesian priors for models that try to fit the model data that fills in the real data we know rather imprecisely where we know it at all but is missing. Yet we rather expect that causal relationships in the climate could be what, ten dimensional? Fifty? Ten thousand?
Do we even have an argument to place a plausible limit on the nontrivial dimensionality of the strange attractor we are checking in projection one or two axes at a time?
Actually, probably we do. Take the climate data in the various dimensions and analyze the noise (or rather, the variance), not the mean. Perhaps when the sun is active, the projective radius of the attractor on a variable’s dimension produces a signal of some sort that isn’t buried in other “noise” that is really causality in still other subspaces of the system. Perhaps the spectrum of this “noise” shifts in tune with the solar data, even though the integral over the noise rarely accumulates into a macroscopic average signal. If one examines in particular things like “discrete” events — volcanic explosions, say — that have a clear spike followed by a smooth decay, perhaps the rates (time constants) of those decays are functional on solar state. Solar state could be leaving its causal fingerprints all over the place, but not having much total effect on global averages of variables taken one at a time.
There is also the second approach. Start looking not for univariate correlation, but for multivariate correlation. Much more difficult, much more subject to the curse of dimensionality. But perhaps useful.
rgb
Peanuts. The string theory struggles with a 10^{500}th dimensionality.
It struggles with much more than that!
rgb ‘ Take the climate data in the various dimensions and analyze the noise (or rather, the variance), not the mean. Perhaps when the sun is active, the projective radius of the attractor on a variable’s dimension produces a signal of some sort that isn’t buried in other “noise” that is really causality in still other subspaces of the system. Perhaps the spectrum of this “noise” shifts in tune with the solar data, even though the integral over the noise rarely accumulates into a macroscopic average signal. ‘
When I look at the NH sea ice anomaly on wuwt’s ref page http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/seaice.anomaly.arctic.png there seems to be distinct phases where the variability shifts from one range to another. With the eye of faith, it’s also possible to see a portion of a cycle centered on 2010, though that would be a lot longer than the solar cycle so perhaps not relevant here. Hope rgb that I haven’t misinterpreted what you said above.
rgb, “Solar state could be leaving its causal fingerprints all over the place, but not having much total effect on global averages of variables taken one at a time.” You are correct.
The solar activity effects are layered and time dependent and sorting them out IS a bear!
Equatorial cloud data can and does correlate strongly with trade winds. The models reproduce this phenomenon rather nicely. Walker cell circulation and disruption factors hugely. And all of that has to do with physics: the shape of our Earth, its inclination towards the Sun, tidal forces due to gravitational pulls of the Sun and moon, and the coriolis effect as Earth spins on its axis. The trick is that one must add inertia and fluid dynamics to the mix which then causes a breakdown in predictability the further away you get from 0 degrees latitude. In other words, you get chaos. Hard to forecast using models because they must simulate chaos as well.
Really hard to forecast using the average of model averages as in the MME mean of CMIP5, because the average of a chaotic trajectory is not a chaotic trajectory at all, and is nearly useless for the purpose it is being used for.
What might be interesting is looking for models that produce trajectories that correspond well with observation, but with chaos and the Kolmogorov scale mismatch, even that is more like looking for chance correlation rather than anything that could conceivably be predictively robust.
rgb
And as I have put in a simpler manner, the climate modelers are trying to model something they imperfectly understand.
Another well argued piece, Willis.
I know it’s not scientific evidence but just anecdotal evidence, but astronomer William Herschel was supposed to have won many a bet on the price of wheat & corn, by counting sunspot numbers. The more sunspots, the lower the price of wheat & corn, the fewer the number, the higher the price, etc. Piers Corbyn hasn’t posted here for sometime but he always talked about the 22 year cycle of full reversal of Solar magnetic field back to it’s previous state as having influence. As they say, absence of evidence is not evidence of absence!
Any ideas?
I’ve analysed his data, the correlation doesn’t exist. I’ve never read that he won bets at it. There’s a good analysis of the question here.
w.
Great stuff Willis,
Something else to consider might be the work of Dr Tim Patterson, professor and director of the Ottawa-Carleton Geoscience Centre, Department of Earth Sciences, Carleton University. He studies sediment, diatom and fish-scale proxy records in British Columbia and postulates longer period marine cycles, all correlating closely with other well-known regular solar variations. In particular, he sees marine productivity cycles that match well with the sun’s 75-90-year “Gleissberg Cycle,” the 200-500-year “Suess Cycle” and the 1,100-1,500-year “Bond Cycle.” The strength of these cycles is seen to vary over time, fading in and out over the millennia. He proposes that the variation in the sun’s brightness over these longer cycles may be many times greater in magnitude than that measured over the short Schwabe (11 yr) cycle and so are seen to impact marine productivity even more significantly. An example of Dr Patterson’s work can be seen here (2007):
http://www.sciencedirect.com/science/article/pii/S0037073804002507
More recently (2012), Kern et al, (High-resolution analysis of upper Miocene lake deposits: Evidence for the influence of Gleissberg-band solar forcing) show similar correlation results for European lake sediments. I don’t know if this is the type of data set and longer period cycle correlation you are interested in studying (one thing is that the resolution is too low to sort out the 11 yr cycle) – two data tables are included at the end of the article linked here: http://www.sciencedirect.com/science/article/pii/S0031018212006748
Thanks, Kirby. What is the evidence for your claimed “1,100-1,500 hundred year cycle” in the sun?
w.
The Bond/Dansgaard-Oeschger semi-periodic climate cyclicity in the North Atlantic area is well established for the Late Pleistocene, less well for the Holocene. There is no clear evidence that it is solar-related.
Willis,
This is proxy evidence of course. The “evidence” is the paleoclimatic interpretation of oxygen-isotope (and other isotopes) records derived from pollen, diatoms, fish scales, etc. from cores extracted from marine sediments. Both isotopic abundance and scale, pollen, and skeletal counts are processed to extract the periodicity. As usual, correlation is not causation.
…
Late Holocene variability in pelagic fish scales and dinoflagellate cysts along the west coast of Vancouver Island, NE Pacific Ocean (2005)
Patterson et al
(abstract): “Correlation of the marine paleoproductivity records observed in Effingham Inlet with solar influenced climate proxy cycles observed in the North Atlantic region indicates that solar forcing at different scales might have influenced the climate in the northeast Pacific as well. In particular an 1100- to 1400-year cycle in regional climate is well represented in the fish productivity proxy and sedimentological record.”
…
Solar Forcing of Climate. 2: Evidence from the Past (2005)
Gerard J. M. Versteegh
(Abstract) “The noisy character and often insufficient temporal resolution of proxy records often exclude the detection of high frequency decadal and bi-decadal cycles. However, on multi-decadal and longer time scales, notably the 90 years Gleisberg, and 200 years Suess cycles in the 10Be and 14C proxy records of solar activity are also well presented in the environmental proxy records. The additional 1500 years Bond cycle may result from interference between centennial-band solar cycles.”
Thanks, Kirby. As you point out, these are paleoproxies. My problem with many such proxies is that the authors almost never establish their claims using modern data.
For example, you show a study of “variability in pelagic fish scales” and dinoflagellate cysts. Well, we have pelagic fish and cysts scales right up to the present. Their study is here … but unfortunately, as near as I can tell they make no attempt to link up pelagic fish scales with solar variations in the modern period. In fact, they say:
That’s great … but where is the putative solar effect? Or are they claiming that somehow the solar variations affect the “nutrient supply and salinity”?
It is for me a recurring problem with many, perhaps most of these kinds of studies. Despite the fact that the processes involved are going on right up to the present, far too often there is no attempt made to see if e.g. the modern levels of fish scales correlate with the modern, known variations in solar.
To me, it’s like calculating the temperature from the tree rings. If tree rights were as good as they claim, able to discern ± 1°C variations in temperature five hundred years ago … then where are the modern records to bear this out?
And if you want a good laugh, take a look at their Figure 2.
Notice the top line, showing the 14C variations (supposedly the index of solar activity) versus what they identify as the warm and cold periods. Now you tell me … do you see any correlation between the putative solar activity level supposedly shown by the 14C variations, and the warm and cool periods?
Also, the variation of the fish scale counts seems only very vaguely connected to the cool and warm periods, and I see virtually no correspondence between the fish scales and the solar activity.
Unfortunately, this is far too common in the paleo world—no bringing it up to the present, and vague handwaving presented as though it were actual evidence. And bizarrely, for their statistical significance calculations they cite that well-known expert on statistics … Michael Mann. Anyone citing M. Mann on statistics desperately needs to go back to basics and start over.
Net result? When I find that kind of nonsense in a study, I toss it in the circular file. And unfortunately, the paleoproxy community has produced reams and reams of those kinds of studies … you need to start from a position of total disbelief. Far too many people google something like “millennial solar cycles” and grab whatever comes up that fits their beliefs. Me, I’m cynical as hell. I start from the assumption that the authors haven’t done a good job, and that they are trying desperately to make something out of nothing … and most of the time, I find that that’s not far from the truth.
Best regards,
w.
I’ve run an Inverse Fourier Analysis on the data from ftp://cdiac.ornl.gov/pub/ndp021/ndp021r1.f13 with MS Excel.
For station 3812 NC ASHEVILLE there’s a period of 10.9 years for every months(januaries, februaries, marchs…) i analyze for the data from 1922 to 1987 since Excel needs a power of 2 (like 64) as number of inputs.
Interesting. Pick another station at random and see if it works. Given the sheer number of stations with data, you may be as right as a broken watch.
Thanks, Pamela, couldn’t have said it better. In addition … how strong is the cycle? I find all kind of cycles down at the “lost in the weeds” zone.
w.
Heh
i don’t understand why you would look for an 11 year cycle, unless you think the cloud of ejecta shielding us dissipates in an 11 year cycle.
Not only that but many flares are directional .. so the shielding will be uneven, and so more effective
if between us and the denser sources of cosmic rays … in addition to the particle density of the solar win generally.
you would need to calculate flight time, distance vs density, and you might have some idea
how long the sun must be quiet for the cloud to thin out ,,, the further out, the more distance
between particles
a very quiet sun after a very active period, if the activity picks up again before cloud from the previous cycle dissipates, would it even show in the record ?
Normal a quiet periods and feeble active periods over more than one cycle, then we might notice.
But would we be able to separate that from PDO or a couple volcanoes or any other factor ?
I doubt you will until we actually go thru another mini ice age … and notice the dang sun has been a wimp for the past 3 cycles …. otherwise your looking historical accounts and no better evidence than correlation
Either that, or get to calculating the particle density gradient of the ejecta cloud in the solar system.
What, you thought it was the “near earth” material that was being talked about ?
Its not the density of the solar wind passing by .. its the density of the material years past us.
Other can feel free to bop me over the head, and point where I am wrong …
But it seems so clear to me, and I assumed others had the same understanding
The density of any solar wind passing by at the moment simply cannot be what
that theory was about,,,, that ,, would be flatly absurd .. that will intercept next to nothing
Its the accumulation, its dissipation rate, is it being replenished, how often, how much etc.
You wont see any 11 year cycle, you would see only the effect of the average over several cycles
How dense is the cloud .. how much material ,,, how long will it be effective if it is not replenished
what is its speed of expansion etc.
If there has been estimates and math on that, Ive missed it, and yet i assumed that was the mechanism
Willis, your hard work is certainly appreciated. You have explored many approaches and the climate does not respond as we would expect. Can you run an analysis of how the earth’s temperature responds to it’s elliptical orbit around the sun so we can have a more robust signal and see how the climate system responds to it. Perhaps, it would give a clue as to what is happening.
But aren’t we looking for anomalies? The solar effect on atmospheric temperature on a cyclic basis has been calculated. The problem is that Earth’s own variability can and does bury that effect. On a clear dry-air day with no reflectance back to space, we know what happens to the solar heating effect. On a fully “veiled” day (clouds, ash, or sulfur dioxide in the atmosphere), we know what happens to that solar heating effect. The problem is that the mercurial Earth slaps that effect all over the place in-between those two extremes in such a chaotic way that the signal cannot be detected anymore.
A number of Willis’s posts have show the possibility that the Earth’s climate system is a closed loop system. The analysis I suggested would help clarify the dynamic structure of a closed loop system.
Willis what you don’t seem to be able to understand.
The historical climatic record shows a good correlation between extremes in solar variability and the climate. The problem for so many is they just do not understand or do not want to understand that at times when solar variability is limited solar/climate relationships are going to become more obscure. Nevertheless post Dalton solar activity had been on a steady overall rise(especially the magnetic component) and overall global temperatures once again move in an upward fashion which was to be expected. Post 2005 solar activity has switched from an active to an inactive state and correspondingly the temperature rise has already ended and soon it will be down.
As I have requested many times show me the data which shows a long period of time when prolonged minimum solar conditions corresponded to a global temperature increase over several years and when prolonged maximum solar conditions corresponded to a fall in global temperatures over several years. I have yet to see such data.
Now going forward in order for solar variability to overcome the noise in the climate system it has to reach certain low value parameters for a sufficient length of time in order to show a significant solar /climate relationship. I have outlined those parameters and believe they will be coming into play as this decade advances.
One condition has already been satisfied which is several years of sub-solar conditions in general now it is just a matter of waiting for the maximum of solar cycle 24 to end and then seeing how deep and long in duration the depth of solar minimum conditions become and see how this translates to the climate.
As far as AGW theory CO2 is increasing year in and year out and temperatures are not responding in an upward fashion ,if that same situation in a reverse way happens with my theory I would admit to being wrong instead of standing by something (agw theory) which will soon be obsolete.
History has shown that the climate of the earth is NOT static and that it has gone abrupt and significant changes from time to time and I say the the most likely route cause for this is the item that drives the climate that being the sun . When something that drives something changes it, it will change what it is driving to a varied degree.
The essence of all my post is that if solar variability is long enough in duration and strong enough in degree of magnitude change it will have a significant impact on the climate but absent this solar/climate connections will be very obscure.
This decade offers a great chance to see if extreme solar conditions will impact the climate.
Finally, someone who actually understands how sun spot activity=climate rises and falls!
I have grown quite irritated by people wanting to prove that the sun, the #1 driver of climate, has little or no effect! Even very slight changes in overall solar activity leads to huge changes on our planet.
And all ice ages end very abruptly due to the sun suddenly becoming more active and the fact that we are now in a cyclical ice age system that began only 3 million years ago shows that something has very much changed over time and not in a good direction and certainly not towards ‘global warming forever’ but the exact opposite.
Perhaps our local star is much less energetic as previously.
My question is where was your climate governor prior to 10000 years ago?
Also the evidence is there in that weak solar/geo magnetic fields will result in a trend (jig saw pattern) to a cooler climate (but to different degrees read below)and this is EXACTLY what is going to take place now going forward.
I want to add this, thresholds, lag times the initial state of the climate(how close to glacial/interglacial conditions climate is), land/ocean arrangements, earth magnetic field strength , phase of Milankovitch Cycles ,random terrestrial events ,concentrations of galactic cosmic rays within 5 light years of earth due to super nova or lack of for example, the fact that the climate is non linear is why many times the solar/climate correlation becomes obscured, and why GIVEN solar variability(with associated primary and secondary effects) will not result in the same GIVEN climate response.
What is needed is for the sun to enter extreme quiet conditions or active conditions to give a more clear cut solar/climate connection which I outlined in my previous post.
The solar criteria I suggested needed to impact the climate to make it more likely to become colder, which I suggest can happen if the prolonged solar minimum continues and becomes more established going forward.
Which you would then ascribe the recent and now paused global warming to something else?
Willis Eschenbach: November 1, 2014
“So once again, I have donned my Don Quijote armor and continued my quest for a ~11-year sunspot-related solar signal in some surface weather dataset. …. [snip] ….. As a result, if there is no sunspot cycle visible in the terrestrial surface weather datasets, then we can assume that none of those phenomena are affecting the dataset.
In addition, I see this as a chance for you guys to definitively show everyone that there actually is a sunspot-related signal that makes it down here to the surface where we live. So I invite you—bring out your best dataset”.
——————-
I don’t have a “data set” to offer, but I will, after a couple hours of thinking about it, offer my critique of your quest to solve a problem, to wit:
Now I will agree there is a per se ”11-year sunspot-related solar cycle” …. but I really don’t think that the max number (quantity) of Sunspots, ….. or the lack thereof, …. that occur during said 11 year cycle can be used as a “signal” for pointing out ….. or determining some surface weather/climate related dataset or event. If said Sunspots only originated on the Sun’s earth-facing equator then their “numbers” might be used as the aforesaid “signal”.
Now the above does not negate the possibility that the presence or absence of Sunspots is in actuality a “signal”, …. and given the observed correlation between Sunspot numbers and climate events on earth , …… then said “signal” is potentially one that only defines a Solar event for which, … that I am aware of, …. there has never been a detailed “dataset” generated for.
My above conjecture is based on the following, to wit:
———–
Excerpted verbiage from: http://en.wikipedia.org/wiki/Sunspot
Sunspots were rarely recorded during the second part of 17th century.
sunspots all but disappeared from the solar surface
observations of aurorae were absent at the same time.
the lack of a solar corona was noted prior to 1715.
a renewal of sunspot cycles starting in about 1700.
The period of low sunspot activity from 1645 to 1717 is known as the “Maunder Minimum”.
===================
Excerpted verbiage from: http://imagine.gsfc.nasa.gov/docs/science/mysteries_l1/corona.html
One of the most puzzling features of the Sun is what has been dubbed “the solar corona problem.” There is a region around the Sun, extending more than one million kilometers from its surface, where the temperature can reach two million degrees. This region, called the solar corona, is where the solar wind originates.
======================
And the “dataset” I made reference to was one that denoted the presence, absence, surface height and/or temperature of the “solar corona”.
If the lack of sunspots is a “signal” of the potential decrease in/of the “solar corona” ….. then it should also be a “shadow signal” of a potential decrease in Solar irradiance that the earth is subjected to, ….. which appears to be confirmed by the surface temperatures during both the Maunder Minimum and the Little Ice Age, … to wit:
http://i.huffpost.com/gen/1578821/thumbs/o-MAUNDER-MINIMUM-570.jpg?1
I now invite ya’ll to “haveattit” with my conjecture.
Wow!
All my life I’ve been told, like many others, that the sunspot cycle influences the earth’s weather.
Willis, rightly so, says we should be able to prove that. Except, despite trying, he can’t. And, apparently, neither can anyone else.
But ideas do not go quietly into that good night and the reasons why are interesting;
a) the earth’s climate is too complicated (i.e. it does influence it, but we can’t see it)
b) the earth’s atmosphere is a giant heat sink that modulates the sunspot influence into the noise level (but, trust me, it does have an effect)
c) you’re looking at it all wrong … you need to be looking at x or y or z
a) and b) smack of belief without data which, given the usual skepticism about climate change, is ironic. c) is just a distraction.
So why can’t we accept that if the impact is almost immeasurable it is likely a third order effect and we can ignore it? Then, maybe, we can stop all the navel gazing over sunspots.
The connection has been found over & over again in study after study for 200 years.
More verbiage without any link to a dataset. Color me unsurprised, milodon.
w.
I most certainly did provide a link. The same one that you have been shown over & over yet continue to ignore.
The precision of the data are a function of being a 50 year average. Is that really all you have to say?
More blatant ignorance of any data that don’t fit your preconceived notion. No surprise there.
That’s a reply to your comment below about the Central England Temperature reconstruction.
At this point the proper reply is that I have shown you dozens of such papers, but you always studiously ignore them, demanding instead just a single one, which you then find some excuse for not considering.
>> says we should be able to prove that
Can you mathematically prove that rain causes wet sidewalks?
We know for a fact that a solar maximum inserts into Earth about an additional 2.75 x10^22 Joules of energy. To put this into perspective, if only 37% of this energy were in the atmosphere, it would be 2 degrees warmer.
And the vagaries of Earth’s atmosphere and churned up or glass smooth ocean can smack that around like a WWF free for all.
Pamela, agreed. It’s like a rogue wave on the ocean. For a long time, some people speculated that they were possible, when by chance, all the various waves at different frequencies might all sum in the same direction. In the last decade, with increased technology, we now know that at any given time, one or more rogue waves is happening somewhere on earth.
I provide my calculation not as a prediction of what will happen at any certain time, but rather, what should we expect as the limits of natural variability. The point is, clearly, nothing extraordinary has yet happened that calls for a new hypothesis.
I’m glad you brought this up, VikingExplorer. Since 1900 the solar constant has a range of ~2.6 W/m^2 from min to max. We know from Trenberth, et al. that only about 17.7% of incoming flux is absorbed (240/1,362 = 0.177), the range of change in downwelling LWR is ~0.46 W/m^2. Multiply that by the canonical value of 0.8 C/W*m^-2 for equilibrium sensitivity and we get 0.37 C for an approximate upper limit temperature change due to solar cycle. Least squares linear regression of TSI only against temperature yields a slope of 0.29 C/W*m^-2, which is both pleasingly close to and predictably lower than the theoretical upper bound. My linear regression model spits out 0.1 C min to max against the residuals of ENSO, AMO, etc., so there’s about 0.2 to 0.3 C of variability I can’t account for — understandable given the noise and observational resolution/uncertainty already much discussed in this thread — as well as the fact that equilibrium and fast transient response to perturbations are two very different animals.
Regardless, my back of napkin theoretical calcs are within striking distance of a linear regression model. Not exactly QED, but a promising result methinks.
https://drive.google.com/file/d/0B1C2T0pQeiaSTndfczY2N0tBWWc/view
Note that prior to 1960, the clear correlation between TSI and residual breaks down, but just eyeballing the graph the effect still jumps out in places.
VikingExplorer, Errata:
Aack, I mixed units. The proper check of model against theory is: 0.29 C/W*m^-2 * 0.46 W/m^2 = 0.1334 C. Regression of TSI against the model residuals gives a range of ~0.1 C, so I’ve got ~0.03 C of variation I cannot explain, a much stronger result.
The SSN [when calibrated correctly] is an excellent measure of the variation of solar output, as is the F10.7 flux. First you have to realize that both measures by their nature are already a 14-day average as they refer to the signal integrated over the visible disk of the Sun. The R-squared measure for the correlation is in excess of 0.97. This is as good as it gets.
We have a good measure of the F10.7 flux back to 1840 [and possible back into the 18th century as well, as the Earth itself is a good instrument for measuring the variable [the EUV flux] causing the F10.7 flux variation: http://www.leif.org/research/Reconstruction-Solar-EUV-Flux.pdf
Willis is asking for data sets, not your hand wringing. So here is my data set with the F10.7 flux
http://www.leif.org/research/F107-Flux-Since-1840.xls
http://www.leif.org/research/F107-Flux-Since-1840.png
And F10.7 flux and SSN follow each other very closely on daily values too:
http://www.leif.org/research/TSI-SORCE-Latest.png
as they should as they are proxies for the same thing: the solar magnetic field.
Here’s your F10.7 data link http://www.ngdc.noaa.gov/stp/space-weather/solar-data/solar-features/solar-radio/noontime-flux/penticton/penticton_observed/listings/listing_drao_noontime-flux-observed_daily.txt
Here’s the daily F10.7 link for recent activity http://www.swpc.noaa.gov/ftpdir/indices/quar_DSD.txt
and further back you can use these http://www.swpc.noaa.gov/ftpdir/indices/old_indices/
This isn’t hand-waving – I have a spreadsheet and model and the calculation for my determination of 120 sfu/day is very easy and straightforward.
However I’d like to see if someone else can figure this out too, given all the hints I’ve dropped over the summer. I even gave Willis another major clue in August. Of course he wants it spoon fed to him.
Here is that clue: http://climate4you.com/images/SunspotsMonthlySIDC%20and%20HadSST3%20GlobalMonthlyTempSince1960%20WithSunspotPeriodNumber.gif
By the way, I’ll give you all even one more clue, and we’ll see who’s smart enough to figure it out: I substitute F10.7 for SIDC number.
When the time comes I’ll relate it for you wrt the revised GSN.
The only real reason I can say this with confidence is I’ve been paying real close attention to the “system” for some time now and have seen that it works out pretty well. Have at it!
You are missing the point: your F10.7 flux is a 14-day running mean.
Typo – Quijote should be Quixote. Much time spent in Spanish?
No, Willis is correct. The original text (1605) used the Quixote spelling. The English spelling probably dates from then, but Spanish has moved on and it is now ‘Quijote’. Although I would not want to make any claims that Spanish is totally unified.
ENSO effects US cloud cover, and El Nino episodes/conditions tend to happen in the same place in most solar cycles. That doesn’t mean that you can reduce it to an 11yr signal though, as the solar wind speed has more than one major drop through the sunspot cycle.
http://www.esrl.noaa.gov/psd/people/tao.zhang/UShydrocycle-9.pdf
The same place? Really? Prove it. And cite “significant”, not “tend” statistics.
The most regular major drop in the solar wind speed through the solar cycle is typically around one year after sunspot minimum. El Nino episodes (conditions*) occurred in 1902, 1913/14, 1925, 1936*, 1946, 1957 (late), 1965, 1976, 1986, 1997, and 2009.
There is a similar frequency of El Nino episodes/conditions at sunspot maxima, exceptions with higher solar wind speed and La Nina episodes at sunspot maximum instead occurred in 1938, 1989, and 2000.
https://www.longpaddock.qld.gov.au/products/australiasvariableclimate/ensoyearclassification.html
There seems to be a lot of belief in the comments that the signal would be obscurred by other aspects of the climate system. I disagree.
The diurnal cycle is obvious and short. The temperature of the atmosphere changes rapidly due to changes in input. Any effects over a yearly cycle would be detectable. I have camped outdoors in many climate zones, one may freeze in the desert at night and fry during the day. a shift of 40 degrees C.
There is a governor in this climate system and it is very powerful, identifying it is a bit of a problem however. I believe that Willis is on the right track with the moisture system ie. clouds.
There’s an emergent phenomena in the afternoon of the solar cycle that’s pretty obvious in this graphic from Joe Bastardi’s recent post:
http://wattsupwiththat.files.wordpress.com/2014/10/clip_image003.png
Perhaps like thunderstorms being feature of the diurnal cycle, El Nino’s might be a feature of the solar cycle; its precise arrival timing and strength depending upon the conditions of the solar cycle.
Please read Joe Bastardi more carefully. The red crosses are carefully cherry-picked El Ninos; most are not depicted.
Thanks, George. That graph is very misleading.
w.
There are two errors, in SC 19 the El Nino was from early 1957, three years after the sunspot minimum, and the last red cross is placed too early, it should be on the 2009/10 El Nino. That major El Nino episodes occur with regularity at the dominant drop in the solar wind speed in most sunspot cycles around one year after each sunspot minimum, is profound evidence for a solar cycle influence on weather and climate.
I’m not sure it’s correct to splice the two cloud datasets this way. The two signals appear to lie nicely on top of each other but amplitudes of the two signals are clearly different – there seems to be a systematic shift where at lower clouds, satellites are picking more cloud cover than humans and the other way around for high cloud cover. By proper recalculating of human observations to satellite measurements, the 1940 rise might decrease in magnitude significantly. Sadly the overlap is very short. It would be probably better to do much more detailed correlation analysis, down to the level of individual stations to have some decent certainity it’s done right.
The way it is done, amplitude of features to the left of the splice is uncomparable to the amplitude of features to the right.
Thanks, Kasuha. I made a conscious decision not to try to adjust the variance (what you call the “amplitude of features”, because I wasn’t really interested in the variance. Instead, I was interested in the trends over time.
Your suggestion, that the reason for the difference in variance, is that humans and satellites see the high and low cloud cover in different ways, is very interesting.
However, I suspect that the reason is actually in the density and regularity of the sampling. An average of individual stations is almost guaranteed to have greater variability than the result of a satellite-style regular sampling of the entire area. From memory, the underlying ISCCP data is something like a 25 kilometer square pixel, and there are a whole lot of those over the US.
I was most encouraged by the fact that the means of the raw ISCCP data and the raw station data in the overlap period (53 months) were within one percent of each other. That plus the high correlation of the two datasets are what convinced me that splicing is justifiable.
All the best,
w.
rgbatduke
November 1, 2014 at 7:36 am
Much more difficult, much more subject to the curse of dimensionality.
Not only in space but in time.
Two of the major indices run on clocks ‘set’ at different frequencies, a bit like two cars, one travelling at 50 and the other 55 mph. Last re-starting point was sometime before or around 1900, there are some indications that both may soon run out of petrol (gasoline), if so I suspect another re-starting point may not be far off.
While we are at ‘pseudoscience’, the three mayor highly respected scientists McCracken, Beer and Steinhilber have moved over to ‘astrology’
http://link.springer.com/article/10.1007%2Fs11207-014-0510-1
Desperate times require desperate measures !
I’ll try another analogy
Two cars on a racing circuit at 150 and 165 mph, no re-start more likely ‘lapping’ over. Unfortunately, good data may not be even one ‘lap’ long.
While we are about solar cycles, SSN is 27 points down at the SIDC’s October count see graph here , making the second SC24’s peak definitely over, is there peak no. 3 coming (?), no idea, but Dr. S some time ago suggested it may be even more.
I hope not. I want those warmists to eat some crow and only the Sun can make that happen!
I hope yes. Longer the cycle, stronger the cooling.
Willis is also showing the lack of any other short term cycles.
Willis Eschenbach wrote shortly after Figure 8:
“I draw no great conclusions from this, except that overall, as the US warmed post-1930, the US cloud coverage rose as well. Go figure …”
I consider this comment as a somewhat cherrypicked sideshow to claim that clouds are a negative feedback to global temperature. However, I notice Figure 8 having cloud reporting peaking in the late 1960s, and most global warming in the time record covered by Figure 8 occurred afterwards.
The way I see it, to the extent Figure 8 shows clouds having a negative feedback on global temperature, I think there is no need to narrow down to a decade – I see that as motivating suspicion of cherrypicking more than it explains a point.
Also, I see this point as a distraction from low rate of visible presence of a ~11-year or ~22-year in surface weather datasets. If your point is that Earth is simultaneously self-regulating and acting randomly, I suggest avoidance of appearance of cherrypicking, of things related to negativity/positivity & chicken/egg of cloud coverage with global (or US) temperature. It appears to me that being more strictly focused on lack of correlation with solar cycles would be better for you to get your point across.
There are different types of clouds no? To me, what’s important is the type of cloud… does the cloud cover give us any indication of that.?