Ian H. Bryce writes at Jo Nova’s website:
The thing that intrigued me about the maximum temperatures is the high peaks, which occur at the peak of the odd solar cycles, and four other times, when we had strong El Nino events. (Most recently, three in four years) It is interesting to note that we did not have the Super EL Nino in 1998!
One wonders when our climate scientists graph global mean temperatures for tens of thousands of stations worldwide, that they “miss the wood for the trees.” I contacted the BOM some time ago about this phenomenon, but I have not had a reply yet. (Surprising?)
Read the entire fascinating article at Jo Nova’s website.
UPDATE: Willis finds some serious problems and posts in comments:
Not sure why I usually end up being the one to rain on the parade, but I’ve accepted my lot in life. Here is the Echuca data plotted against the peaks of the solar cycles, as measured by sunspot counts.
A couple things of note. First, he has misidentified the Cycle 11 peak, it happens earlier. Second, he is very vague about the timing of the cycles. Yes, the high years occurred during those cycles, but if we look at the actual peak year of each cycle, some happen two years before the peak temperatures, some three years before, some four years before, and some show no relation at all to the peak temperatures.
Sorry … but that’s the real data, and the sunspot/temperature correlation doesn’t hold up in the slightest.
SOURCES:
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Clearly something is wrong with that graph. It doesn’t trend upward with a violent uptick post 1980.
I heard them on TV- several times, different stations- the scientists were all saying that the sun doesn’t have anything to do with our Climate Change problems, that it’s all because of the CO2 from the industrialized free world. and that we have to shut down outr industry and do it now, while there’s still a chance.
I’m supposed to believe some lady with a blog over MSNBC and CNN?
This may be an artifact related to spurious correlation expected to occur when there are thousands of stations to look at. It can also be expected to occur when two disconnected oscillations occasionally coincide, much like the windshield wipers do on the bus when controlled by separate motors. These coincidental correlations can stay together for many decades, depending on the oscillation of each, but may not at all be connected to the same motor.
Seems that there is a correlation of peaks to odd number cycles.
That is not surprising. The sunspot cycles flip polarity of the magnetic fields alternately, so it looks to me as though the underlying cycle is actually around 22 years.
What we observe as the 11 year cycle is an artiface of the existence, or not, of sunspots. There is no way of expressing a “negative” sunspot number, so they all become positive. The cycle clearly has a positive and a negative half and the temperature peaks correlate to one side.
I wonder what the underlying mechanism is.
Does the Northern Hemisphere respond to the even numbered sunspot cycles? 😉
I think I’ve seen previous work about possible linkage between the Earth’s and Sun’s magnetic fields, and there are ways in which the even/odd cycles may be relevant. The two magnetic fields are in similar alignments during every other cycle — but the Sun’s magnetic fields are rather peculiar.
I’m curious to see the atmospheric temperature before, during, and after each.
Did the global temp increase similar to 1998? How much of the heat stuck?
Echuca is a good example of a continuous temperature record from 1881 until the present, in a town which has experienced absolutely no UHI effect. In 1881 Echuca was probably about the same size and density as it is today. This strongly suggests that the typical industrial era increase in average temperature seen in most cities temperature record`is an artefact of increasing size, density and energy use of the city.
I’ve expected for a long time that both solar polarity and intensity may be a factor in warming.
Of course, correlation doesn’t mean causation. Perhaps changes in high energy cosmic rays affect some of the dynamics in the sun too.
Not sure why I usually end up being the one to rain on the parade, but I’ve accepted my lot in life. Here is the Echuca data plotted against the peaks of the solar cycles, as measured by sunspot counts.
A couple things of note. First, he has misidentified the Cycle 11 peak, it happens earlier. Second, he is very vague about the timing of the cycles. Yes, the high years occurred during those cycles, but if we look at the actual peak year of each cycle, some happen two years before the peak temperatures, some three years before, some four years before, one is four years after, and some show no relation at all to the peak temperatures.
Sorry … but that’s the real data, and the sunspot/temperature correlation doesn’t hold up in the slightest.
SOURCES:
Sunspots
Temperature
What could be interesting is what will happen with SC 25. L&P have forecast a maximum Wolf number of 7.
I’m with Willis:
The sunspot cycle peaks do not match up. The first two cycles 11/13 are >7 years out, the later ones 2-3 years.
Someone is making stuff up.
I’m impressed that Echuca had an aerodrome in 1881, more than 20 years before the first aeroplane was invented. Now that’s forward thinking for you. Perhaps the attitude was “Build it and they will come”
@braddles, prior to that the station was in town. Moving weather stations to airports at the dawn of aviation was a fairly common practice. The name then gets changed for the station.
What we observe as the 11 year cycle is an artiface of the existence, or not, of sunspots. There is no way of expressing a “negative” sunspot number, so they all become positive. The cycle clearly has a positive and a negative half and the temperature peaks correlate to one side.
That would be an interesting plot, reversing to negative the even cycle counts. Shouldn’t be too hard either, the hardest part being to decide what is the proper date when each cycle begins and ends.
(repeating a comment made at Jo Nova’s) Using the KNMI temperature data, looking at Irkutsk in Siberia, you get a marked 11 year cycle through the 1800’s. Doesnt show in most of the 1900’s. William Connelly made same comment as Willis Eschenbach. Might be a coincidence. Irkutsk is remarkable in its nearness to a giant freshwater lake (Baikal) Not so for Echuca.
I have also noticed some passing resemblance between the ENSO and changes in the geomagnetic field, possibly just coincidence:
http://www.vukcevic.talktalk.net/ENSO-dB.htm
Willis I appreciate your sharp eye. Now please would you make it even sharper, just a little bit, and answer my honest question:
Is there a smaller but still statistically plausible correlation between the odd-numbered solar maxima and the temp. highs? It looks to me as if there might be… with a time offset… I’m wondering if there could be a solar effect mediated by ocean shifts like El Nino which are only semi-predictable and likely to be lagging solar changes (like seasons lag solstice/equinox).
”
Luther Wu says:
June 25, 2012 at 11:47 am
I heard them on TV- several times, different stations- the scientists were all saying that the sun doesn’t have anything to do with our Climate Change problems, that it’s all because of the CO2 from the industrialized free world. and that we have to shut down outr industry and do it now, while there’s still a chance.
I’m supposed to believe some lady with a blog over MSNBC and CNN?
”
Luther,
you think on average that there’s a significant difference in the quality of information from some lady with a blog versus msnbc or cnn? Selecting some blogger gives you around a 50/50 chance that the data is better than or the same (or worse) than msnbc or cnn. I wouldn’t advise putting any strong amount of faith in either, but at least with the ‘lady with the blog’ there is a chance of something with some accuracy. I doubt you’d have a similar chance with either msnbc or cnn. Unfortunately, they’re not consistent enough in their errors to just take what they say and invert it to get the correct information. However, if you could bet odds on it, betting against anything that msnbc or cnn says will likely make more money for you.
Ignorance and total lack of competence seems to outweigh the serious political bias one finds there.
Reagan almost said it right. “trust but verify”. If it’s important, it should simply be “don’t trust, verify!”
cba says:
June 25, 2012 at 2:14 pm
“…”
____________________
Dang!
Forgot my sarc tag.
Lucy Skywalker says:
June 25, 2012 at 1:50 pm
First, I haven’t a clue what you mean by “mediated by” in your post above.
Second, we’re already fairly deep into data mining. We start with one picked temperature dataset and no idea how many other datasets were discarded to find that one.
Next, we have a vague association with some solar cycle peaks but not with others.
On top of that we have a claimed association with some but not all of the El Ninos, including a total miss on the big 1998 El Nino.
Now, you want to add in a lag based on “a solar effect mediated by ocean shifts”, whatever you might mean by that.
The problem is, you have asked for a “statistically plausible correlation”. If we look at one dataset and find 95% statistical significance (p-value less than 0.5), that actually means something.
But if we look at as few as a dozen datasets, the odds are is almost a 50/50 that we will find a result with what looks exactly like 95% statistical significance purely by chance.
This is a recurring problem in climate science, where people dig through reams and reams of data, find one dataset where some relationship or trend has a p-value less than 0.05, and declare it is statistically significant … NOT.
w.
Willis, I appreciate your attention to detail, as I do with Leif when it comes to anything to do with solar cycles.
One thing that struck me was that the idea of a solar peak as a point in time is a very artificial concept. In reality, solar activity is spread out, with periods of higher activity and periods of lower activity.
I don’t have the tools to do it myself at the moment, but I would be interested in seeing what the graph you did looked liked if, rather than a single point for each solar cycle peak, it was plotted with a bar indicating the range of peak activity for that solar cycle. eg. The period of time when solar activity was within X of the peak (using whatever value for X – fixed amount or percentage – that seemed appropriate). After all, some solar cycles have broader peaks than others.
I suspect it still won’t show a reliable relationship, but it would be “fairer” than picking a single date as the “peak” for each solar cycle and looking for a relationship with just that single date.
Peaks in single years are unimportant. Average maximum temperature is highly correlated to the number of sunny days throughout the year which is random variable, no wonder the record goes up and down by large amount. Still, there is pretty nice pattern at least between 1920 and 2000 where the record appears to somewhat follow the sun activity. I guess Svensmark would like it.
Every generation is rediscovering things that had been done before. Echuca is a good example of the Hale Cycle. There are plenty of references to the 22 year cycle in rainfall. Here’s a few:
From: http://ccb.colorado.edu/lanina/report/oh.html
Before 1974, the 2-3 year cycle was dominant in precipitation variations. However, this cycle diminished when the transition occurred and then switched to a longer 3-5 year cycle. It is also interesting that the 22-year cycle becomes dominant when the signal of 2-3 year cycle diminishes as a result of the climate shift.
http://www.geo.umass.edu/climate/theses/waple-thesis.pdf
The Hale (22-year) cycle has for some time been invoked as a possible reason for a bidecadal drought rhythm in the Southwest USA and Great Plains. It is possible that this apparent cycle of drought could be due to internal variability of the ocean-atmosphere system, but as Cook et al. (1997) indicated, it is also not easy to eliminate the sun (and moon in this case) from the causal
hypotheses. It appears from Cook et al.’s findings, that since at least 1800, the lunar cycle
(18.6 years) and the Hale cycle interact to modulate the drought cycle.
From: http://www.springerlink.com/content/a2geaywxlu5dan8w/
The investigation results of the monthly precipitation and Palmer’s drought severity index (PDSI) data in three of the midwestern states – Illinois, Indiana and Ohio – show that two periodic components, the 20 to 22-year Hale cycle signal and the other component with periods between 16.9 and 13.5 years, are identified.
Also see the chapter “The Twenty-Two-Year Drought Cycle in the Western United States” in “The Role of the Sun in Climate Change” by Hoyt and Schatten. It starts on page 138.
Graeme W says:
June 25, 2012 at 2:48 pm
Thanks, Graeme. Actually, the peak of the sunspot cycle is generally pretty pronounced. But that’s not why I’m going to pass on your request. The problem is that your request takes us into what I call the “kinda sorta” territory.
If you plot the entire solar cycle, then you can say “Well, the temperature kinda sorta has a relationship to a carefully selected band which includes times when the solar activity is within X of the peak” … that way lies madness. The relationship shown in my graph varies from three years before the temperature peak to four years after the peak, and includes both temperature peaks with no corresponding solar activity, as well as solar peaks with no corresponding temperature activity … in other words, no relationship.
I used to drill water wells for my daily bread, and one thing I learned was, a wise man knows when he’s digging a dry hole … and this one here is dry as a bone.
w.
Kasuha says:
June 25, 2012 at 3:36 pm
Thanks, Kasuha. You clearly illustrate the problem with the human eye, which is trained and honed to detect patterns … so much so that it easily and regularly detects patterns that aren’t even there.
To repeat myself, the relationship shown in my graph varies from the solar cycle maximum being three years before the temperature peak to four years after the temperature peak, and includes both temperature peaks with no corresponding solar activity, as well as solar peaks with no corresponding temperature activity … in other words, not a “pretty nice pattern”, but in fact, no pattern at all.
All the best,
w.
PS—Our eyes are trained to detect patterns so that we can a) detect food, and b) detect predators. In both cases, the penalty for false positives (thinking you see a tiger or a mango when you don’t) is much, much smaller than the penalty for false negatives (not noticing the tiger or the mango that is there).
As a result, we see lots of stuff that, when we look carefully, turns out to be an illusion … but that’s way, way better than not seeing the tiger. On the flip side, however, it is a constant problem when looking for patterns in scientific data, which inter alia is why we have statistics.