For those of you who want a thread on this paper (which I don’t have time to read right now)”
http://nzclimatescience.net/images/PDFs/alexander2707.pdf
This study is based on the numerical analysis of the properties of routinely observed hydrometeorological data which in South Africa alone is collected at a rate of more than half a million station days per year, with some records approaching 100 continuous years in length.
The analysis of this data demonstrates an unequivocal synchronous linkage between these processes in South Africa and elsewhere, and solar activity. This confirms observations and reports by others in many countries during the past 150 years. It is also shown with a high degree of assurance that there is a synchronous linkage between the statistically significant, 21-year periodicity in these processes and the acceleration and deceleration of the sun as it moves through galactic space.
Despite a diligent search, no evidence could be found of trends in the data that could be attributed to human activities.
It is essential that this information be accommodated in water resource development and operation procedures in the years ahead.
Climate Audit 
68 Comments
Alas, an analysis that may allow true predictions of the future, without obscure computre modeling. This is a truly great paper, as far as I can tell (don’t know if the statistics are up to par, but I’ll wager they are).
This is a great example of someone doing genuine work to solve a genuine problem and in doing so discovers information which has wider ramifications rather than those who devise a problem and then work(?) to find reasons for its excistence. Fascinating reading.
The paper considers the 22 year double sunspot cycle to be important, not the 11 year cycle. In figure 1, alternative sun spot cycles are multiplied by -1, and it is this modified series that the authors claim fits with climate (though autocorrelations are not considered). We are being asked to believe that one sunspot maxima causes drought, the next brings rainfall. What mechanism could explain this? The paper doesn’t seem to have any suggestions, and it cannot be easily reconciled with the hypothesis, suggested by Nir Shaviv and others, that cosmic rays influence climate.
At Lake Victoria, the authors find that the 11 year cycle correlates with lake level (not the double cycle). So now we need a mechanism that fits the double cycle in S. Africa, but the single cycle in Central Africa. The portion of the record where the correlation between sunspot numbers and lake level collapses is not plotted.
I’m not going to discuss the final portion of the paper as Stellvia has already commented on unthreaded on the weak support for the notion that there is a link between sunspot activity and solar motions.
[snip- this is unacceptable language]
3:
I think there is likely to be a readily explainable mechanism for the alternating effects of the successive 11 year cycles. The sun’s magnetic polarity reverses itself every 11 years, around the peak of the sunspot cycle. So for 11 years, the sun’s polarity is opposite of that of the earth…and for 11 years, the sun’s polarity matches that of earth. Therefore, it is to be expected that the magnetic fields of the earth and the sun will interact differently when the polarities match than when they are reversed. In fact, one would probably expect to see the effect most emphasized at the poles of the earth, and least at the equator. Given that Lake Victoria is very near the equator, it should be afected by changes from both hemispheres, while south africa would be affected mostly by effects occurring at the south pole of the earth. Looking at the anomaly between southern and northern hemisphere spots on the sun during each cycle and comparing the effects in the north and south hemisphere on earth, taking into account the relative polarities of the two bodies will likely shed a lot of light on the mechanism by which the single and double cycles impact earth’s climate.
BCL,
Three examples for you…if I give them will you shut up and argue science instead of attacking the personalities involved?
1. George Taylor, Oregon State Climatologist – http://www.kgw.com/news-local/stories/kgw_020607_news_taylor_title.59f5d04a.html
Governor wants to take his title away because he won’t join the collective and swear allegiance to the consensus.
2. Jasper Kirby, Leader of CERN’s CLOUD experiment – http://www.canada.com/nationalpost/news/archives/story.html?id=975f250d-ca5d-4f40-b687-a1672ed1f684
Dr. Kirby abruptly lost funding for the CLOUD project after daring to express his doubts about AGW.
3. Any TV meteorologist who wants to keep their AMS certification – http://www.weather.com/blog/weather/8_11392.html
Dr. Heidi Cullen is calling for the AMS to decertify any meteorologist who refuses to join the collective and swear allegiance to the consensus.
So there you go. One guy is battling for his title as state meteorologist because he refused to submit. Jasper Kirby lost 10 years worth of work on his project, and Heidi Cullen wants to take away certifications from anybody who doesn;t see things her way. Three concrete examples of places where people who didn’t bow before the consensus have lost their title, research funding, or are bing threatened with losing scientific certifications.
So now you shut up about messengers and tell me what is wrong with the science BCL. You never seem to want to talk about the science in any of your little trolls here. Why is that? Are you afraid of an honest debate about science at a place other than RC, where comments contrary to the consensus are actually allowed to remain for all to read instead of being quickly censored so that nobody who reads the threads are aware that there is any dissent?
#4&9 — bcl, you’re just making a varient on the argument from authority, the variation being that it’s stoked with youthful arrogance rather than intellectual arrogance.
Whether young or old, scientists argue from fact and theory. Not from scorn and dismissal. Ideas stand or fall on their own merits, not on the age of the person making them. If you have a science-based case to offer, then let’s see it. Arguing as you are just makes you look shallow and foolish. and makes your case look empty
#6 — Bill, I suspect the 1/r^2 fall-off of magnetic fields is going to make any direct solar-Earth magnetic interaction very weak. But you have an interesting point. The magnetic field lines of the sun presumably direct the trajectories of charged solar wind particles and cosmic rays. When the solar field reverses, the handedness of the charged particle spiral will reverse in each hemisphere and particles entering the Earth magnetic field will go off on a different tangent. Maybe by 90 degrees? Some crackerjack solar physicist like Doug Hoyt could figure out if that has any influence on the way that clouds might be induced. But if it did, one might see the effect oscillating between northern and southern hemispheres as the solar field changed polarities.
4: thanks for the snip.
#9: And this is EXACTLY what is shown in a study I found. Can’t find the ref. right now, but I have it and will look. Maybe Nir Shaviv. The AGW crowd is extremely afraid of the possibility of a Solar connection, so they fight it very strongly. Esp. IPCC.
Regarding intimidation of AGW dissenters, here’s what Lindzen wrote in the WSJ in 2006:
“So how is it that we don’t have more scientists speaking up about this junk science? It’s my belief that many scientists have been cowed not merely by money but by fear. An example: Earlier this year, Texas Rep. Joe Barton issued letters to paleoclimatologist Michael Mann and some of his co-authors seeking the details behind a taxpayer-funded analysis that claimed the 1990s were likely the warmest decade and 1998 the warmest year in the last millennium. Mr. Barton’s concern was based on the fact that the IPCC had singled out Mr. Mann’s work as a means to encourage policy makers to take action. And they did so before his work could be replicated and tested–a task made difficult because Mr. Mann, a key IPCC author, had refused to release the details for analysis. The scientific community’s defense of Mr. Mann was, nonetheless, immediate and harsh. The president of the National Academy of Sciences–as well as the American Meteorological Society and the American Geophysical Union–formally protested, saying that Rep. Barton’s singling out of a scientist’s work smacked of intimidation.
“All of which starkly contrasts to the silence of the scientific community when anti-alarmists were in the crosshairs of then-Sen. Al Gore. In 1992, he ran two congressional hearings during which he tried to bully dissenting scientists, including myself, into changing our views and supporting his climate alarmism. Nor did the scientific community complain when Mr. Gore, as vice president, tried to enlist Ted Koppel in a witch hunt to discredit anti-alarmist scientists–a request that Mr. Koppel deemed publicly inappropriate. And they were mum when subsequent articles and books by Ross Gelbspan libelously labeled scientists who differed with Mr. Gore as stooges of the fossil-fuel industry.
“Sadly, this is only the tip of a non-melting iceberg. In Europe, Henk Tennekes was dismissed as research director of the Royal Dutch Meteorological Society after questioning the scientific underpinnings of global warming. Aksel Winn-Nielsen, former director of the U.N.’s World Meteorological Organization, was tarred by Bert Bolin, first head of the IPCC, as a tool of the coal industry for questioning climate alarmism. Respected Italian professors Alfonso Sutera and Antonio Speranza disappeared from the debate in 1991, apparently losing climate-research funding for raising questions.
“And then there are the peculiar standards in place in scientific journals for articles submitted by those who raise questions about accepted climate wisdom. At Science and Nature, such papers are commonly refused without review as being without interest. However, even when such papers are published, standards shift. When I, with some colleagues at NASA, attempted to determine how clouds behave under varying temperatures, we discovered what we called an “Iris Effect,” wherein upper-level cirrus clouds contracted with increased temperature, providing a very strong negative climate feedback sufficient to greatly reduce the response to increasing CO2. Normally, criticism of papers appears in the form of letters to the journal to which the original authors can respond immediately. However, in this case (and others) a flurry of hastily prepared papers appeared, claiming errors in our study, with our responses delayed months and longer. The delay permitted our paper to be commonly referred to as “discredited.” Indeed, there is a strange reluctance to actually find out how climate really behaves. In 2003, when the draft of the U.S. National Climate Plan urged a high priority for improving our knowledge of climate sensitivity, the National Research Council instead urged support to look at the impacts of the warming–not whether it would actually happen.
“Alarm rather than genuine scientific curiosity, it appears, is essential to maintaining funding. And only the most senior scientists today can stand up against this alarmist gale, and defy the iron triangle of climate scientists, advocates and policymakers.”
re: #12,
Then there’s the treatment of Bjorn Lomborg in Denmark.
The study finds a strong correlation between water levels and sunspot numbers. But the correlation is a short term correlation – there is little to no correlation in the long term trends.
For example, there is no long term trend in Lake Victoria’s levels from 1900 to 1940 when solar activity showed long term increase. Any short or long term correlation breaks down between 1930 to 1970. Next, to obtain correlation over 1968 to 2005, they filter out a 29mm per year trend. They don’t explain why there’s been a long term trend of falling water levels over the last 37 years while there’s been no long term solar trend.
In fact, all the case studies show short term correlation with solar cycles but no long term correlation with decadal solar trends. There’s so much noise due to tributary inflows, outflows, sluicing, rainfall and evaporation that while short term correlations with the solar cycle are useful, determining or finding meaning in long term trends is problematic. In short, measuring water levels is a roundabout way of determing the sun’s effect on long term global warming.
A more direct method would be to observe the correlation between solar activity and global temperatures. In 2005, Sami Solanki at the Max Planck Insitute compared solar activity & temperatures over the past 1150 years and found temperatures closely correlate to solar activity. When sunspot activity was low during the Maunder Minimum in the 1600’s or the Dalton Minimum in the 1800’s, the earth went through ‘small ice ages’. The sun has been unusually hot in the last century – solar output rose dramatically in the early 20th century accompanied by a sharp rise in global temperatures.
However, Solanki also found the correlation between solar activity and global temperatures ended around 1975. At that point, temperatures started rising while solar activity stayed level. This led him to conclude “during these last 30 years the solar total irradiance, solar UV irradiance and cosmic ray flux has not shown any significant secular trend, so that at least this most recent warming episode must have another source.”
#14 Nir Shaviv disagrees:
“Solar activity has been increasing over the 20th century. Thus, we expect warming from the reduced flux of cosmic rays. Moreover, since the cosmic ray flux actually had a small increase between the 1940’s and 1970’s (as can be seen in the ion chamber data), this mechanism also naturally explains the global temperature decrease which took place during the same period. Using historic variations in climate and the cosmic ray flux, one can actually quantify empirically the relation between cosmic ray flux variations and global temperature change, and estimate the solar contribution to the 20th century warming. This contribution comes out to be 0.5⯰.2 C out of the observed 0.6⯰.2 C global warming.”
“There are two reasons why the temperature should rise from the 1970s. First, there is a decrease in the average cosmic ray flux*. If we look at the average of each cycle there is an increase in the average cosmic ray flux until about the cycle of 1970, and then a decrease in the following two cycles. The last cycle was not as strong, so the average CRF increased. This can explain why the temperature stopped warming from around 2000.
Second, one has to realize that the temperature response of Earth’s climate is a ‘low pass filter’ due to the high heat capacity of the Oceans. This implies that:
The temperature variations over the 11 year cycle are highly damped (but t hey are there at a level of 0.1 deg).
There is a delay time in the system’s response. This means that the 11-year cycle will lag the solar forcing (and it does by 1-2 years). Over the centennial time scale, the Sun’s activity significantly increased until the middle of the century, then it slightly decreased and somewhat increased from the 1970’s with a peak in 2004. If you pass this behavior through the climate “low pass filter”, you will find that because of Earth’s heat capacity, the temperature at 2000 should be higher than the temperature in 1950’s even if the decrease until the 1970’s is similar to the increase afterwards.”
re: #13,
“least this most recent warming episode must have another source.” This doesn’t mean we should jump on the AGW Bandwagon. In addition, the acceptance of the Maunder Minimum and Dalton Minimum would rule out AGW; anything but stable temperature is considered high heresy.
Looking at the Technical paper, I’m mightily impressed; though not totally surprised ‘it just seems to make sense’. Personally, I like this idea as it reinforces the solar flux postulate and explains solar spots.
Their’s more work needed but it’s still important stuff.
#6
The changes in the polarity of the sun’s magnetic field between successive sunspot cycles has no appreciable impact on the cosmic ray flux.
#14,
Before you can declare that there is no correlation between solar activity and earth temperatures over the last 30 years, you must first remove the various contaminations of the temperature record. Starting with UHI. Which despite the claims of certain papers is significant, and growing.
All the “unacceptable language” has been put in one place:
http://bigcitylib.blogspot.com/2007/06/team-denier.html#links
Re: 3 Actually the link between sunspot activity and solar motion is not a new idea. It was first proposed in 1965 by Paul Jose. The article can be found here http://tinyurl.com/33qgj7
According to google scholar the article has been cited 69 times, and there are other contemporary articles that promote a similar theory.
The article’s conclusion is:
“The relationships set forth here imply that certain dynamic forces exerted on teh sun by the motions of the planets are the cause of the sunspot activity. This is supported by the 178+ yr periods in the sun’s cycles.”
#14 JC:
If you read the text, you might see that an explanation of the reason for the drop in water level is indeed given on p. 39:
The reason seems to be ASGM (Anthropogenic Sluice Gate Management), not AGW.
RomanM
I quickly read through the Alexander paper last night and this morning and I must say that I was disappointed by the handling of the data by the authors.
The nation of South Africa suffers from weather extremes in terms of rainfall and the government apparently has a major interest, therefore, in the ability to predict and prepare for drought and flood conditions. I believe this paper is the result of scientists/engineers being tasked with finding a model on which those predictions can be rationalized. In my mind this would tend to promote the torturing of the data to make it tell something and that is what, in my view, these authors have done.
They initially “invent” a double sunspot cycle to fit what they consider is an irrefutable 21 year hydrometeorological cycle and then in conclusion talk about a sun/earth 21 year periodicity of the orbiting as influenced by the solar systems center of mass.
While casual viewing can allow for some periodicity in the hydrological factors it is not at all clear how a prior they were able to establish a 21 year cycle — or at least by viewing the variations in flow of the Vaal River. There are several year to year extreme transitions that go from dry to wet and some of these correspond to beginning of the 21 year periods.
The 21 year cycles never showed up in the data from the 20 evaporation sites; for rainfall they see the 21 year periodicity to a 95 % CL in 18 of 93 sites and claim it is present in 67 and absent in 8; for 28 river flow sites they see the periodicity at a 95% CL in 7 sites, claim it is present in 12 and absent in 5 and not determinable? in 4; finally for flood peak maxima they see the periodicity at a 95% CL in 4 of 17 sites , claim it is present in 7, absent in 2 and not determinable? in 4. In my view if one takes all these findings together they make a weak case for the 21 year periodicity that they claim is irrefutable.
I was thoroughly confused when they started with the double sunspot cycle (DSC) and ended with the solar system center of mass periodicity (SSCM) argument, since my casual reading did not detect a link between DSC and SSCM.
The authors make a major claim that the satellite measurements (and those used by the IPCC) of variations in solar radiation are in error because they do not agree with those that can be readily calculated using the SSCM model. That claim I would think could be easily verified or refuted.
The authors’ performance here fits too closely what I have seen of those who have data snooped investing strategies in finding relationships without a good explanation for potential cause and effect. I, therefore, have very major reservations with their analyses and results.
re 22.
Thanks Kenneth. Personally, I’d like to see just as vigerous discussion of this paper
as we had of Parkers and hope folks take you up on the merits.
as you wrote
when I read that I was expecting such a comparison. It would seem to me that a direct comparison
of a model (SSCM) with recorded data, would be most instructive.
Re #17
RichardT,
I suppose it depends on what you mean by no appreciable impact but the image you reference clearly shows a significant change in the waveform shape between 11-year cycles. The cosmic ray peaks of 1965 and 1987 have a sharp sawtooth shape whereas the peaks of 1976 and 1998 are so broad that it’s hard to call them peaks. I don’t know yet what the implications are for observed climate but I don’t think one can dismiss the solar dipole orientation quite so easily.
Regards,
Earle
One interesting thing about this article is that they arrive at the same conclusion as Landscheidt, that the determining variables of Earth climate are variations in the centre of mass of the planetary system and in the consequent variations in the angular momentum of the Sun. Landscheidt was able to accurately forecast the El Nino/La Nina cycle using a model based on this hypothesis.
#22, 23
There is also some weakness with the relationship in the “analysis” of the relationship between the world maximum floods and their cycle timing:
The p-values for testing a 10:11 split (as indicated in Table 6) versus systematic difference (direction not predicted) are .153 and (predicting more likely during first cycle) .079. This is not exactly compelling support for a link.
RomanM
re: #26,
BTW, I just flipped a coin 40 times. Results 17 heads, 23 tails. A 16-24 split is nothing special. If it were 6-34 I’d take notice. BTW, one thing people often forget when deciding whether something is worth noticing is to allow for which results would be noticed. In this case either 16:24 or 24:16 would be equally interesting. So after calculating the odds of 24 or more tails out of 40 you’d have to multiply the odds by 2 to see just how unlikely you’d take it to be.
#14
I suppose measuring tree-rings is a much more direct method?
Re: 22 “They initially “invent” a double sunspot cycle ”
Actually, they didn’t invent a 22 year double sunspot cycle. It was first described by the astronomer Hale in the early 20th century, in fact it is known as the ‘Hale cycle.’
http://en.wikipedia.org/wiki/Hale_cycle
Re #26, #27
I apologize for not having read the paper in detail, but would it not help our understanding better to know how many floods occurred in each of the cycles? The authors are missing out on an opportunity to bolster that observation by not making it clear, or else perhaps by lumping them all together they are implying a stronger correlation than really exists.
Richard,
The cosmic ray flux is WAY more complicated than that graph can possibly show. What is traditionally measured as “cosmic rays” are neutorns…which as that graph shows are directly and strongly tied to the 11 year solar cycle. However, the “galactic cosmic rays” make up a very small but important fraction of the total flux of high energy particles. Changes in the quantities of those particles would be totally swamped by the 11 year period in total cosmic ray flux. The GCRs that Svensmark and others postule as impacting cloud formation are a specific form with a specific energy level called muons. The energy level of the particle is specifically what allows it to interact with the atmosphere in a way that can impact cloud formation.
Here is a paper that talks about the relationship between the earth’s magnetic field and the ability of muons to penetrate all the way to the surface. http://www.research.ibm.com/journal/rd/421/ziegler.html
The key finding of that paper is that the ability of muons to reach the lower atmosphere and the surface of the earth is greatly affected by the earth’s magnetic field, which is in turn buffeted by the solar wind. It is well known that the interaction between the sun’s magnetic field and the earth’s magnetic field (known as the interplanetary magnetic field of IMF) determines how strongly the solar wind affects the earth’s magnetic field. In particular, the Bz component of the IMF leads to strong aurora activity at the poles when it is strongly south oriented.
This paper demonstrates that there are period of time where for an unknown reason, the one polarity of the BZ or the other dominates. (http://www.agu.org/pubs/crossref/2005/2004GL021110.shtml). They speculate that it is the shift of heliomagnetic equator away from the heliographic equator that causes the preferential polarity.
In short, we know very little about how and why changes in sunspots occur. We do know that the interaction between the sun’s magnetic field and the earth’s magnetic field has a significant impact on how the solar wind affects the earth, particularly at the oles. And finally, we know that the incoming muon energy is greatly affected by the earth’s magnetic field, which is particularly sensitive to the solar wind…especially at the poles.. So posting a record of total cosmic ray flux (not muon flux or galactic cosmic ray flux) from a single station at high altitude and mid latitude, and claiming that because it doesn;t have an obvious visual signal, that there must be no influence of solar polarity on climate is a bit of an overstatement of what the graph really shows. Specifically, the highest change in flux created by differences in polarity would be expected to be apparent at polar locations, while the paper’s findings with regard to lake Victoria suggests that there would be little to no difference at locations closer to the equator.
If anyone wants to check annual rainfall data for cycles, this Australian BoM site lets you quickly download annual rain numbers 1900-2006 for a variety of Australian regions and for each state. At first glance there seems to be diverse patterns.
http://www.bom.gov.au/cgi-bin/silo/reg/cli_chg/timeseries.cgi
#30 Earle:
I didn’t chase down the catalogue of world’s major floods, so I can’t speak to the general case. The 40 “greatest floods” were tabulated in the paper’s Table 6. They referred to the 24 out of 40 as meaningful so that is the issue I addressed. I also did a simple plot of number of floods versus cycle year. Year 6 with six floods sticks out a bit, but there is no apparent pattern otherwise. It looks pretty much trendless across the sequence. I suspect more information would be needed to try show some sort of global flooding – sun cycle relationship.
RomanM
Alexander et al. very kindly include the river discharge data from Vaal Dam in their paper. This makes it easy to audit their claim of a 21 year cycle in this data. So here is 15 lines of R code that replicates their result in Figure 1, and then shows it not robust.
vaal=c(39,246,42,66,44,83,142,40,36,24,170,131,87,225,59,202,112,131,54,185,353,87,66,58,57,33,100,33,60,100,45,181,80,277,188,69,75,105,50,68,58,149,27,175,31,35,60,52,102,23,112,295,247,123,122,31,63,62,19,12,79,30,36,46,208,165,65,59,13,26,92,17,464)
vaal=ts(vaal, start=1923)
x11()
plot(vaal, type="h")
x11();par(mfrow=c(2,2))
hist(vaal)#Highly skewed data
qqnorm(vaal)
qqline(vaal)
#Log transformation of the data to give it a ~Gaussian distribution
hist(log(vaal))
qqnorm(log(vaal))
qqline(log(vaal))
x11();par(mfrow=c(2,1))
acf(vaal, lag.max=25)# Identical to Figure 1. Minimally significant at lag 21.
acf(log(vaal), lag.max=25)# Transformed data not significant at lag 21.
And I imagine this was one of their better examples.
#34 After a brief look at the data, I agree with your point about the skewness. I would be a little careful in looking only at the acf, however, in determining significance at lag 21. For example, a moving average model on the log transformed variable gives significant lags at 11 and 21, although this might be spurious…:
. arima lnflow, ma(11,21)
(setting optimization to BHHH)
Iteration 0: log likelihood = -100.99338
Iteration 1: log likelihood = -96.728884
Iteration 2: log likelihood = -94.704042
Iteration 3: log likelihood = -93.810029
Iteration 4: log likelihood = -92.941927
(switching optimization to BFGS)
Iteration 5: log likelihood = -92.855037
Iteration 6: log likelihood = -92.723507
Iteration 7: log likelihood = -92.707627
Iteration 8: log likelihood = -92.705058
Iteration 9: log likelihood = -92.704982
Iteration 10: log likelihood = -92.704982
ARIMA regression
Sample: 1923 to 1995 Number of obs = 73
Wald chi2(2) = 23.62
Log likelihood = -92.70498 Prob > chi2 = 0.0000
——————————————————————————
| OPG
lnflow | Coef. Std. Err. z P>|z| [95% Conf. Interval]
————-+—————————————————————-
lnflow |
_cons | 4.289158 .076358 56.17 0.000 4.139499 4.438817
————-+—————————————————————-
ARMA |
ma |
L11. | -.8329472 .2045405 -4.07 0.000 -1.233839 -.4320551
L21. | -.8208876 .1913873 -4.29 0.000 -1.196 -.4457753
————-+—————————————————————-
/sigma | .6520827 .091964 7.09 0.000 .4718365 .8323289
——————————————————————————
#32 – Interesting.
I’ve been looking at river flow data since this thread started. I had downloaded annual flow for the Ohio River at Metropolis, Illinois here (picked this one randomly since it had a history back to 1929): http://waterdata.usgs.gov/ky/nwis/current/?type=flow
Calculated the average annual flow and the accumulated deviation from the average. Plotted in Excel with the sunspot cycle and ran an 11-year average trendline through the sunspot cycle. That was interesting itself.
I just did the same thing with the Australian Region precipitation which Warwick Hughes provided the link for. From the looks of the plot, the rainfall in Australia affects the flow in the Ohio River [sarcasm]. Both the precipitation and the Ohio River flow have similar curves to the 11-year average sunspot cycle. I’d share the plots, but don’t have a way to host the images.
#21 RomanM, good point. My initial interpretation was the sluicing was before 1968 but you’re right, it may have been long term sluicing from 1968 to 2005. But that just reinforces my original statement that the study is useful for short term correlations but too many factors invalidate it as useful for long term correlations.
# 36, pk
Try http://www.esnips.com/ to host your images. I’d like to see your plots.
#38, Thanks John G. Bell, let’s see if this works.
http://www.esnips.com/web/PrecipSunspots
#39
It works, but it works better if you use the Img button in the Quicktags bar. Position the text cursor, click on the Img button, put in the URL click OK, enter the description and click OK and you’re done. See (I hope):
Australian precip vs Ohio River flow – cumulative deviations from the average:
This is a test.
#40, Okay, that didn’t work. Following the link seems to be best.
Never works when you want it to. Must be Murphy’s law. The hyperlink is there because if you right click on the description text and select View Image, it sends you to the right place. Groan.
I don’t know if esnips is the best place to host images. Photobucket.com worked for me.
It appears that river flow is a better correlation to sunspots than individual gauges because the river flow itself is a massive averager of total precipitation over a wide area.
IMO The surface temperature network has too much man made noise in it to pick up the solar signal out of the noisy data set. The atmospheric water cycle however seems to have no trouble divulging the solar variance.
More studies are needed, I’m going to contact my friend Jim Goodridge, the formar California State Climatologist who is doing boatloads of precip studies right now to see if we can pick up the same trends in California’s rivers such as the Sacramento and San Joaquin rivers.
#45
Anthony:
I am all for more data from different sources, but isn’t river flow data also subject to anthropogenic factors such as changes in land use and irrigation?
#45, yes land use changes affect river flows. I did a study of the Chattahoochee River in Georgia a few years back, modeling the river flow over about a 50 mile stretch. I didn’t quantify, but you could see changes from different watersheds depending on the amount of urbanization. Hydrographs in highly urbanized areas were sharper (runoff was faster) and there was less infiltration leading to lower low flows. However, the total flow was largely driven by precipitation. I don’t think it would have as significant an affect as looking for tenths of a degree while measuring temperature over asphalt.
I’m not saying there is a correlation between sunspots and precipitation, but so far I’m finding this interesting and plan on looking at a few more rivers as I get time.
Re #46
bernie,
You’re quite right. Anything we measure directly is subject to anthropogenic factors. On a global scale we have influenced the mount of particulate, trace gasses, etc. in the air. On regional levels we have the significant changes in land use contributing to changes in humidity, albedo, and heat island effect.
Obviously our local climates are much more complex than the simple anthropogenic global warming index (AGWI) promoted by the IPCC. So we should be measuring all those things we can about these climates and all those things we can about our land use and other influences on the local climates. It may be that there are so many confounding factors that water flow is one data source that we cannot reliably use. On the other hand it may be possible to allow for those influences and either correct (measure upstream consumption and discharge) for them or mitigate them (do not use data from specific drainages).
My guess is that it will be a lot easier to identify and quantify those confounding factors for water flow than for tree ring density and width. And it is very easy to specify exactly what areas influenced the observed measurement. There will be a few rare exceptions, but as a rule the flow is indicative of prior precipitation within a sharply defined boundary.
An exception that comes to mind is the Gulkana Glacier in the Alaska range, near the Richardson Highway. Glacial melt currently runs into the Delta river and eventually to the Yukon River. When the USGS mapped it in the 1950s it drained into the Gulkana River and eventually via the Copper River to the Prince William Sound. On the scale of measurable river flow in the Yukon and Copper rivers this glacial melt may be insignificant, but it is one of the natural confounding factors that one should take into account.
I’m doing a lot of arm-waving here, but my bottom line is that you can’t tell if there will be problems without looking at all the data.
RE46,47.48 Bernie you are correct in pointing that out, pk your point about the total flow being driven by precip is well taken too. And Earle, yes we must look at all the data. The good news is I don’t think we’ll find BBQ’s next to river flow gauges to confound the measurement. But there may be something else…who knows.
I’m going to use the “beer” analogy here. Beer is a “rental drink”. While you may sweat some of it off (evaporation) the input flow (drinking it) and the output flow (well, you know) tend to correlate pretty well no matter how many pretzels and peanuts you eat (urbanization temporarily inhibiting flow).
Re: #34 and #35
My look at the Vall River data made it difficult to see a 21 year cycle and as RichardT has noted I would suspect that that case appeared as an example in the text because it was one of the better ones. Could you both describe in more detail what your analyses are showing in posts #34 and #35? I understand essentially what you are doing but could you give the probabilities (i.e. better/worse than 5% for happening by chance). They are a number of people who are here to learn.
Regarding my post #36, I’ve put a few more plots of river flows/sunspots up if anyone is interested.
http://www.esnips.com/web/PrecipSunspots
Let’s try this again. This time I right clicked on the graph and selected copy image location. It’s a different URL from the link to this file box. It probably still won’t work.
Hurrah! That was the medium size. Here’s the original size:
Thanks DeWitt Payne.
So far, I’ve selected only major rivers (5,000 to 10,000 annual average CFS or more) with decently long history, trying to get to about 1930 or earlier. I’ve tried to include different regions of the US as much as I can so far.
51. There seems to be a rough relationship between flow and sunspot activity in your graphs, but every once in awhile it “misses a beat.” Maybe something, like el Nino overrides everything? The Colorado River one may not make sense, since it is below a dam.
#51, “it misses a beat”: one might expect that depending on what region they’re in. Are they influenced by activity in the Pacific Ocean which could have a different frequency than the sunspots. Re: the Colorado River: Does it deliver it’s precipitation/river flow to other watersheds – affecting the inflow/outflow relationship?
I was just think if the sun spot cycle is 11 years whats the solar year of jupiter; mmmm.
http://personal.eunet.fi/pp/tilmari/tilmari6.htm#221
well buger me with a hockey stick.
Didn’t we have a planetary alignment a few years ago?
Professor Alexander has sent me the South African hydrological data used in his paper, and given me permission to distribute these data to anybody who would like to analyse them. I have written R code to import these data that I can make available.
I interrupt normal programming to announce that I have received answers to questions, generated in discussion at the ClimateAudit blog, directed to D.E. Parker concerning his study on the Urban Heat Island effect, and its effect (or lack thereof) on the perception of global warming through land-based temperature measurements.
You can find his responses at:
http://www.climateaudit.org/?p=1718#comment-119294 ,
starting at entry #386.
The more sceptical comments in the thread above seem to be borne out by a re-analysis of Alexander’s data, which can be found here: http://www.wrc.org.za/downloads/waterwheel/mar-apr07/Letter%20p%2011-13.pdf
But is a perfectly fine procedure if your name is Lockwood and you are using it to disprove a link between galactic cosmic rays and global temperature. Goose. Gander. Sauce.
#61
Thank you for posting this. Very useful.
#62
There is nothing wrong with smoothing data. Indeed, there are often good theoretical reasons for doing so, but the proviso is that the analyst must acknowledge that there is a reduction in number of degrees of freedom, or there is an inflated risk of a Type 1 error.
I would agree with the criticism of Alexander’s paper in this link. I think the flaws are made much more evident in this paper because the authors have yet to master some of the more obfuscating techniques used by Mann and his cohorts.
Is there an 18 year solar cycle? I’ve run a wavelet analysis on the log-transformed data Vaal Dam inflow data, using this online wavelet program.
At the 90% (but not 95%) confidence interval (against a red noise background) there is a cycle at about 18 years, nothing at 21 years.
I disagree with the hypothesis of Nir Shaviv because I found the opposite in the short term. I don’t know the reason; perhaps the difference obeys to the class of cosmic rays that we have considered on our plots. I think Dr. Shaviv is considering the normal oscillations of the incoming CR without taking into account the anomalies found by the spaceships Voyager 1 and 2. I have read a bit from Alexander’s paper and it is almost the same phenomena that I found with respect to the Interstellar Cosmic Rays incoming to Earth from the bow shock of the Solar System. I published some graphs regarding my observations made from 2002 to 2005. I got the ICR data from Voyagers I and II. The anomalies of the incoming ICR are exactly the anomalies of temperature registered on Earth throughout the same period. The graphs and annotations are here.
You can check the accuracy of the information from my article with any book on Astrophysics. I recommend you also to read the further reading, incorrectly linked in my article like References.
A spectral analysis of the data listed in the paper for the Vaal River gives the following periods in years:
18.34, 10.35, 6.49, 4.46, 2.37
Alexander uses only the differences between the minima on his page http://pielkeclimatesci.wordpress.com/2010/05/20/climate-change-the-west-vs-the-rest-by-will-alexander/ whereas maxima occur in 1924 and 1996, 4 cycles later, giving 18 years rather than 21 years.
The period of the lunar nodal cycle is 18.6 years, which has a strong influence on tides, both ocean and atmospheric. It is much more likely that this is the cause of the longest cycle rather than the Sun. However the cycle detected as 10.35 years might well be the solar cycle.