Many have sought, but few have found, a connection between the solar cycle and global temperature. Such a connection only makes sense; solar irradiance varies with the solar cycle, being greater when the sunspot cycle is at maximum and less when it’s at minimum. And solar irradiance means energy coming into the climate system; more energy input should cause a temperature increase. Yet the change in solar irradiance throughout the solar cycle is small, and the resultant change in climate forcing is smaller still, so the impact is likely to be hard to find. In addition, the response to changes in forcing is likely to lag behind the forcing itself, and when that forcing is periodic the lag can strongly dampen the response. So, the temperature response to solar-cycle variations is likely to be very small indeed, which may explain why it’s so hard to pin down.
Let’s take a look at some data, and see what we see. For global temperature I generally prefer NASA GISS, but that only goes back to 1880. The Hadley Centre data goes all the way back to 1850, so for global temperature anomaly we’ll use the HadCRUT3v data set:
Solar irradiance has only been measured with accuracy during the satellite era. Some claim there are secular changes over time, some claim there aren’t, but since we’re looking for solar cycle changes we can exploit the relationship between sunspot counts and solar irradiance. So as a proxy for solar cycle forcing, we’ll use sunspot numbers:
Although these plots display annual averages, for analysis purposes I’ll work with monthly data from 1850 to the present.
The dominant feature of sunspot numbers is the roughly 11-year Schwabe cycle. This is clear from a Fourier transform of sunspot numbers:
The tallest peak indicates the solar cycle, with a period of 10.8 yr. This is actually the mean period from 1850 to the present; the actual cycle length isn’t constant. The peaks very near the main peak are indicative of the non-constancy of the solar cycle; they indicate both frequency and amplitude modulation of the dominant cycle.
The peak at the far left simply indicates that the data show long-term changes. Fourier (and all period) analysis often indicates peaks at low frequencies (long periods) which don’t indicate real periodicity, just changes on long time scales. I’ve spent a good portion of my career trying to educate scientists on this matter; the tendency to interpret low-frequency peaks as genuine periods is one of the most frequent mistakes made by actual research scientists (and I’ve seen some extreme examples). In this case, the very low frequency peak corresponds to a period of about 150 years — given that we only have a little over 150 years of data, deducing the existence of a 150-year periodic change is obvious nonsense.
This doesn’t prove there’s no 150-year period — but there’s no real evidence of it either. The sunspot data only demonstrate a single periodic behavior, the 11-year Schwabe cycle; there’s no sign of the 22-year Hale cycle, or of longer cycles, in sunspot observations since 1750. There is actually some evidence of a period near 8.3 years, which is of highly variable amplitude, but that requires further investigation before considering it established.
Let’s see what we get from a Fourier transform of the HadCRUT3v data:
There are some tall peaks at very low frequencies, but again they’re indicative of characteristic time scales associated with long-term changes rather than genuine periodicity. There’s no sign of any significant peak at a period near 10.8 yr. In fact there’s no sign of any genuinely period behavior in the temperature data.
We can look a little closer if we remove the long-term trend from the data. Doing so will suppress much (but not all) of the response which is due to characteristic time scales in the trend, and make it easier to detect genuine periodicity if it’s there. Smoothing the data on a 30-year time scale and analyzing the residuals, here’s what I got:
Once again, none of the peaks is really tall enough to constitute significant evidence of periodic behavior. The peak near period 22 yr is actually due to remaining longer-term behavior, and doesn’t pass proper statistical evaluation anyway. The peak near frequency 0.011 cycle/yr also fails significance, and even if it passed it’s at the wrong period to match the solar cycle; at period 9.14 yr, if it represented genuine periodic behavior then in the time span of this data it would have gone through two more cycles than the sun! There is a small peak at about the right period, but it’s nowhere near statistical significance, and if it does represent solar-cycle response it indicates a semi-amplitude of only 0.03 deg.C, or full amplitude of a mere 0.06 deg.C. So we still haven’t found real evidence of a solar-cycle signature, and if what we have found does represent it, its amplitude appears to be very small indeed.
The solar cycle itself is not strictly periodic because each cycle has a slightly different period and amplitude. Perhaps the solar-cycle response shows similar behavior? Period and amplitude modulation will suppress the response of a Fourier transform to this kind of changing periodic behavior. This suggests applying a wavelet analysis to the data. For the sunspot data we get this:
We see a clear response in the expected frequency range of about 0.09 cycle/yr (period about 11 yr), consistently throughout the entire time span, as would be expected from a genuinely cyclic phenomenon. The changing frequency of the “ridge” indicates the changing period of the sunspot cycle, which we can estimate thus:
The sunspot cycle period has indeed shown variation in the last century and a half. Does the temperature data show signs of periodicity with similar variations?
The brief excursions in the wavelet transform don’t actually represent periodicity, just temporary episodes of near-periodic behavior. There are also sizeable spans of time without even a hint of periodic or “characteristic time scale” behavior in the region we’re scanning. And of course, if we do interpret the temporary episodes as periodic fluctuation, the indicated periods don’t match the solar-cycle variation:
We really don’t see evidence of periodic fluctuation in global temperature data which in any way resembles the solar cycle. Of course there are other ways to test for a solar-cycle/temperature connection. We can seek to correlate temperature with sunspot counts themselves, allowing for a time lag for the solar variations to take effect. The result indicates no significant correlation. We can remove the impact of volcanic eruptions from the temperature data, then repeat the above analyses. Again, the result gives no evidence of a solar-cycle/temperature connection. It’s starting to look as though whatever connection may exist, it’s so weak that the noise level of the data makes it difficult or impossible to confirm. I have heard from Urs Neu that if one removes the impact of volcanic eruptions and the el Nino southern oscillation (ENSO) from global temperature data, then it is possible to detect a weak response of global temperature to the solar cycle. But I haven’t run the numbers, so I can’t confirm or deny it.
The most plausible detection of a solar-cycle response in temperature data is by Camp & Tung (2007, Geophysical Research Letters, 34, L14703). At the outset they mention the failure of previous efforts:
There have been thousands of reports over two hundred years of regional climate responses to the 11-year variations of solar radiation, ranging from cycles of Nile River flows, African droughts, to temperature measurements at various selected stations, but a coherent global signal at the surface has not yet been established statistically [Hoyt and Schatten, 1997; Pittock, 1978].
Camp & Tung look at the geographic pattern of temperature changes, and how that differs between times of high solar activity, and times of low activity. In this way they delineate a “solar cycle geographic pattern” for temperature change. They then correlate the observed geographic pattern of temperature changes with solar irradiance (using the reconstruction of Lean et al.) to estimate the amplitude of the solar-cycle response, and to test for statistical significance of the relationship. They determine that it is indeed significant, and estimate the global average temperature response to the solar cycle as having an amplitude of 0.18 deg.C.
I’m not convinced by the Camp & Tung result; there are many complicated steps in their analysis, which increases the level of uncertainty. Also, their analysis uses only data from 1959 to 2004, and although it may be based on a superior representation of solar output it also restricts the available data to less than 50 years, which raises further doubts. In addition, I may have identified a flaw in their method which artificially inflates the indicated solar-cycle response. However, I haven’t analyzed the situation in sufficient detail to be sure. I’m not going to outline my suspicion in detail because the idea belongs in the peer-reviewed literature, not on a blog, and because I should communicate my concern to Camp & Tung and ask for their opinion before stating any details anyway. I only mention it to explain why I’m not convinced that their result is correct. I’ll emphasize that by no means am I convinced that it’s wrong.
If Camp & Tung’s result stands the test of time, it’ll represent the first genuine confirmation of a solar-cycle signature in global temperature. The amplitude is not very big, but is bigger than the limits indicated by the various period analyses we’ve applied, which raises the question why it hasn’t been detected by other methods.
The extreme difficulty finding a solar-cycle signal in global temperature indicates the low level of the response, which is in accord with theoretical estimates of what the response should be. The many efforts, and considerable ingenuity, applied to the search, are a testament to the tenacity and cleverness inherent in scientific research. As future data, both for global temperature and especially for solar output, continue to accumulate, we can expect ultimately to determine the response of global average temperature to the solar cycle with confidence. I look forward to both the journey, and the destination.
106 responses so far ↓
kim // April 5, 2008 at 8:16 pm |
Leif Svalgaard claims that solar cycles have alternate rounded and pointed peaks of TSI. Someone else has shown that cosmic rays correlate with clouds in one cycle, but not in the next. The PDO has a cycle approximately 33 years long and alternates cooling and warming phases. At three solar cycles per PDO cycle, one PDO cycle with two correlating solar cycles and one without, is followed by a PDO cycle with two non-correlating solar cycles and one with correlation. If more clouds cool the earth, and fewer warm it, we have the explanation for the last century of cooling and warming PDO cycles alternating approximately every 33 years.
This is simple. Perhaps too simple. But William of Ockham would take a look at it, anyway.
=============================
The Tuatara // April 5, 2008 at 9:31 pm |
Very clear. Thanks HB, most useful.
David B. Benson // April 5, 2008 at 9:43 pm |
Tamino — Once again, very clear. Thank you!
nanny_govt_sucks // April 5, 2008 at 9:56 pm |
Maybe because you’re looking at sunspot numbers alone. In the second 11 years of the 22 year Hale cycle the sun has reversed its polarity. A good analysis of solar effects on Earth’s climate would take this into account and look at the 22-year cycle, and odd/even 11-year solar cycles. Also what Kim said.
David B. Benson // April 5, 2008 at 10:34 pm |
kim & nanny — I quote from a Michael Tobis blog post:
“In a recent talk at the University of Texas, Z. Liu of Wisconsin described some experiments he did with two different climate models in tracking down something called a Pacific Multidecadal Oscillation, a sort of long-term wobble in the climate of the Pacific. Because there is so much interest in the equatorial zone in the Pacific (due to El Nino) most attention to this phenomenon had been in the tropics, but Liu observed that there was no known process in the tropics of the right time scale. There is such a process in the far north Pacific, called ‘baroclinic Rossby waves’ the mathematics of which I won’t trouble you with.
“Liu wanted to test his idea that these very slow waves are crucial to the PMO. He found some climate models that display a PMO in their statistics, and put a giant simulated sponge in the North Pacific that would leave the rest of the system as unimpeded as possible but would suppress the baroclinic Rossby wave. Sure enough, the PMO went away in the model. A couple of further experiments confirmed that putting the sponge elsewhere had little effect. He repeated this with two distinct model codes.
“This constitutes strong support for Liu’s hypothesis.”
Eli Rabett // April 5, 2008 at 11:59 pm |
What part of there are no 11 or 22 year cycles in the global temperature record don’t folks understand?
Hank Roberts // April 6, 2008 at 12:27 am |
Many decades ago, when television was black and white and sometimes funny, a comic once said, “Somewhere on Earth, every six seconds, a woman has a baby. Our task is to hunt her down and stop her.”
Similarly, one (perhaps even two, or three) can argue that the heat coming from the sun must arrive at just one surface station, somewhere — and be measurable there — even though it diffuses out and gets lost in the global average.
I’m sure someone’s trying to find that location. It’d be where the hyperthermic magnetospherical tubicular flux tube arrives on the planet.
Rattus Norvegicus // April 6, 2008 at 12:47 am |
Hank, would that be Roswell, NM?
Leif Svalgaard // April 6, 2008 at 2:21 am |
Kim, you have misunderstood something. The ’rounded’ and ‘peaked’ cycles are in cosmic rays [and the cause is well understood]. The solar cycles themselves do not show such behavior.
Hank Roberts // April 6, 2008 at 3:01 am |
So, Tamino, in the spirit of anticipatory whackiness, you could also look at whether there’s a solar cycle detectable in the changes of the earth’s rotational speed; maybe there’s just a variable heat input from electromagnetic field strength braking and accelerating the planet, eh?
I just made this exquisite new theory explaining the sun’s influence up, but here’s the proof that the idea is worth someone (not me, I don’t have a clue how) doing the heavy lifting to disprove it:
http://www.nature.com/news/2008/080403/full/news.2008.717.html
http://www.agu.org/cgi-bin/SFgate/SFgate?&listenv=table&multiple=1&range=1&directget=1&application=sm07&database=%2Fdata%2Fepubs%2Fwais%2Findexes%2Fsm07%2Fsm07&maxhits=200&=%22GP54A-07%22
Hank Roberts // April 6, 2008 at 3:05 am |
Oh, if these folks are still operating, they ought to have the data needed:
http://dx.doi.org/10.1016/0273-1177(93)90217-Y
The International Earth Rotation Service: Current results for research on Earth rotation and reference frames
Central Bureau of IERS, Observatoire de Paris, 61 Avenue de l’Observatoire, 75014, Paris, France
John Cook // April 6, 2008 at 3:21 am |
Another paper that looks for a solar cycle signal in global temperatures is Scafetta 2005 (http://www.fel.duke.edu/%7Escafetta/pdf/2005GL023849.pdf). She finds a cycle but with about half the amplitude of Camp & Tung. But the striking feature of her analysis is how the temperature cycle goes out of phase with the solar cycle the further back you go. So if there is a correlation between solar cycles and global temperature over the past few decades, perhaps it’s more coincidence than causation.
[Response: Scafetta & West have done a number of papers on sun-climate connections, which in my opinion represent nothing more than shoddy work. Some of it has been critiqued on RealClimate, here, here, and here.]
kim // April 6, 2008 at 3:36 am |
OK, Leif, thanks for the explanation. Still might not rounded and peaked cycles of cosmic rays explain one solar cycle correlating with clouds, and the next, not? I know there isn’t a lot of data about solar cycles alternating with and without a correlation between clouds and cosmic rays, but if that continues, it might explain a slight change in solar energy causing a greater change in climate. What might cause one solar cycle to have a such a correlation, and the next to not have it?
====================================
cohenite // April 6, 2008 at 3:41 am |
Solar climate influence works by proxy; for instance, with less sunspot activity there is less UV and X-ray flux; since these are preferentially absorbed in the upper atmosphere with consequent heating, with less, the upper atmosphere cools, shrinks and causes an increase in the jet-stream and movement of polar air towards the equator; a concentrated thermal uplift occurs with an increase in clouds and general cooling as Spencer has observed.
The response time of the proxies is considerable. Scafetta uses a 7.5 period and achieves good correlation with temp over both an 11 and 22 year cycle. The 22 year cycle may just be an amplification of the 11 year cycle with the latitudinal positioning of the sunspots, variations in solar equitorial rotation and the Gnevyshev and Ohl rule all mitigating any clear-cut, linear and immediate correlation between sunspot activity and temp.
In respect of the cosmic ray proxy, Svensmark and Friis-Christensen, in a recent rebuttal of Lockwood and Frohlick’s and Sloan and Wolfenden’s critiques of cosmic rays, note that there is an excellent correlation between cosmic ray activity and tropospheric air temp and ocean sub-surface temp but not with global surface temp.
[Response: Scafetta and West demonstrate sloppy statistics more than anything else, and the so-called "rebuttal" of S&F is no such thing.]
DrCarbon // April 6, 2008 at 4:02 am |
Given that all reasonable folk seem to think the sun-temp is a weaker forcing than well mixed GHGs, whence the data presented in figures like this one from the notorious OISM project:
http://www.oism.org/pproject/Slides/img2.html
(Recall that RCs intent to have this paper cooperatively debunked on a wiki never really took off, so I’m curious…)
[Response: The OISM is nothing more than a propaganda organization. That's why they've plotted data for solar irradiance which is simply not plausible. Compare their plot of solar output to the estimate of Lean et al., or the more recent reconstruction of Svalgaard.]
Tom Woods // April 6, 2008 at 7:41 am |
I think trying to find a connection between Schwabe/Hale cycles and climate is fundamentally worng. Sure there are some factors, such as an increase in TSI, within these cycles that lend to a small fraction of climate forcing, but one also has to consider various other aspects of solar activity when trying to ‘find a connection to climate’.
Solar wind, proton flux, electron flux, x-ray flux, interplanetary magnetic field and dozens of others.
Until one can devise a model which takes all this into account and correlate the findings to real world observations I don’t think anyone can either prove or disprove any connection between solar-climate changes.
Gavin's Pussycat // April 6, 2008 at 9:18 am |
One thing I don’t understand. In Ray Pierrehumbert’s article last december on RC he offers the following graph:
http://www.realclimate.org/BD3.jpg
The red curve, TGlobe, is referred to as “Jones’s curve” and apparently depicts global mean temperatures (is this different from HadCRUT?).
Anyway, in this curve the 11-year periodicity is clearly visible even without Fourier, and the amplitude looks about right by Camp and Tung: 0.18 deg C peak-to-trough, give or take some. What did I miss?
BTW problems with proper interpretation of the discrete Fourier can also happen on the high-frequency end. The method assumes periodicity, i.e. circularity, and a function containing a trend will present a skip between end and start values in the circular interpretation. Fourier will translate this skip into hi-freq “noise” (essentially the transform of a step function) overlaying the result.
Taking residuals from a 30-year moving average helps, but not 100%.
[Response: I don't know the origin of that data or that graph. But you say "in this curve the 11-year periodicity is clearly visible even without Fourier." I think that's yet another case of fooling yourself with an unquantified visual impression. I've published quite a few papers on Fourier analysis, particularly the statistical properties of same, and I don't see it.]
Frank O'Dwyer // April 6, 2008 at 9:43 am |
“They determine that it is indeed significant, and estimate the global average temperature response to the solar cycle as having an amplitude of 0.18 deg.C.”
Supposing this result stands, what would be the impact on the understanding of CO2’s effect?
Leif Svalgaard // April 6, 2008 at 10:50 am |
Hank, the Earth’s rotation is influenced by climate and weather [and even a tiny, tiny bit by solar activity] so hold the sarcasm.
Leif Svalgaard // April 6, 2008 at 12:10 pm |
Kim: as I remarked over at CA: They did not find a difference in response between the two cycles. They found no response in one of the cycles, suggesting that what was seen in the first cycle was spurious. This often happens, people see something, but it doesn’t hold up when more data comes in. Possibly as simple as that.
cohenite // April 6, 2008 at 1:10 pm |
Thanks for that link to your reply to the Svensmark & Friis-Christensen paper. Friis-Cristensen is currently in Australia; I’ll try and get some feedback.
You rely on GISTEMP and HadCRU for temp data; why not UAH and RSS? In any event much is made about the volcanic influence on the data, what about the 1976 ‘Pacific Event’ as detailed in this (unreviewed) article?
http://mclean.ch/climate/Aust_temps_alt_view.pdf
[Response: Please no more links to crackpot theories.]
Hank Roberts // April 6, 2008 at 2:49 pm |
Leif, not meant as sarcasm, the new report suggesting a mechanism via conductive minerals is interesting. My point is any external torque ought to show up in that database and be susceptible to analysis. I keep seeing the notion of mysterious forces, whether planetary alignments or solar magnetic connections, advanced — if they exist they ought to show up in something as closely watched as Earth’s rotation, wouldn’t they?
Tom Woods // April 6, 2008 at 2:58 pm |
Tamino,
The PDO and its influence on climate is not a crackpot theory. You of all people should know this.
[Response: I didn't say the PDO was a crackpot theory. But what's been done with it in the link, is a crackpot theory.]
Leif Svalgaard // April 6, 2008 at 3:14 pm |
Hank, agree on avoided the voodoo, but the simple heating of oceans and atmosphere [and their thermal expansion] changes the moment of inertia of the Earth and hence the LOD. Ice-skater and stretching out of arms… The thermosphere expands with sun’s magnetic field influence and FUV, etc. As the precision is driven higher and higher, more and more of these tiny effects become observable.
Hank Roberts // April 6, 2008 at 4:52 pm |
Leif, thanks. Is what’s known sufficient to answer the kind of comment made above?
http://tamino.wordpress.com/2008/04/05/stalking-the-elusive-solar-cycletemperature-connection/#comment-16477
Leif Svalgaard // April 6, 2008 at 5:03 pm |
Hank, to wit the thousands of papers and the utter confusion that reigns with claims and counterclaims, it seems to me that we either do not have the data or there is nothing to it. The situation is similar to the debate a hundred years ago whether the sun was responsible for geomagnetic activity. At that time, they didn’t have the data [Solar wind, IMF and ionospheric data] and so couldn’t decide. It turned out that there was something to it. I don’t know what we are missing this time or if there is something to it. We just to keep at it [but can do without the politics].
Tom Woods // April 6, 2008 at 9:46 pm |
Tamino,
Can’t argue with you there. The way that ‘paper’ was presented could be considered crackpot. I don’t know how anyone could still hold the belief that increased GHG’s can have no effect and oceanic oscillations holds all answers. I’m sure it’s a combination of, well, everything. Finding those key percentages is what is proving tricky.
Joel Shore // April 7, 2008 at 2:06 am |
In response to Frank O’Dwyer’s question (”Supposing this result stands, what would be the impact on the understanding of CO2’s effect?”): In a sequel to their Geophysical Letters paper available here http://www.amath.washington.edu/research/articles/Tung/journals/solar-jgr.pdf , Camp and Tung argue that the implication is that the climate sensitivity estimated by the IPCC seems to be correct…i.e., that their result implies a climate sensitivity for doubling of CO2 of at least 2.3 C. (Implicit in this conclusion, I think, is the assumption is that it is the actual change in total solar irradiance that is causing the temperature variation and not some other forcing such as some cosmic ray effects or whatever.)
Leif Svalgaard // April 7, 2008 at 2:02 pm |
Joel: Camp and Tung note that “there is a recurrent warming of the earth by the solar cycle. The periodic nature of the phenomenon allows the use of more sophisticated signal processing methods to establish the reality of the signal”. There is also a strictly periodic [hence known forcing] variation during each year of 90 W/m2 due to the eccentricity of Earth’s orbit. This variation, being strictly periodic, and 100 times larger than the solar cycle variation should allow for even more sophisticated analysis. It would seem to me that unless we can model and understand the response to this very large signal, it is premature to look for the much weaker solar cycle signal.
[Response: Just for perspective -- the eccentricity variation in the intensity of sunlight is about 90 W/m^2, but the variation in climate forcing is only about 17 W/m^2. Which is still a substantial amount, and still about 100 times as large as the variation in climate forcing due to the solar cycle. Also, high-frequency periodic forcings tend to be damped more than low-frequency periodic forcings, and this is a very-high-frequency variation.]
Leif Svalgaard // April 7, 2008 at 4:20 pm |
very-high-frequency? what is the time step in climate models?
[Response: I was thinking of the solar cycle itself being reasonably high-frequency compared to the relaxation time for the climate system (about 30 yrs according to climate models). The frequency of the annual cycle is 11 times greater. I don't know what the time step of climate models is.]
Leif Svalgaard // April 7, 2008 at 6:01 pm |
and clarify the 17 W/m2. I would get 90/4 = 22.5, then maybe multiply by 0.85 to account for removal of UV in the stratosphere [under the assumption that that absorption has no effect], leaving 19 W/m2. What else?
[Response: There's also albedo, about 0.3, so I multiply the 22.5 by 0.7. I'd allow the UV stratosphere absorption; it does enter the earth system.
But any way you look at it, it's still about 100 times as big as the solar-cycle variation.]
Leif Svalgaard // April 7, 2008 at 6:55 pm |
time step in models: in order to say that the annual variation dampens out, the time step must be much less than a year and the dampening must follow naturally from the model calculations, i.e. the model must show that dampening occurs. If we just assume that the annual variation MUST dampen out and artificially do so [build it into the model] the models are no good IMHO. One may note that the solar cycle is not a smoothly progressing thing but consists of a sequence of impulse or episodes also of 1-2 years duration. A good example is cycle 23 and cycle 14 [http://www.dxlc.com/solar/cycl14.html].
[Response: I wasn't referring to model simulations, only to the fact that in general systems with a relaxation time longer than the period of the cyclic forcing tend to exhibit more damping of higher-frequency forcings. As for models, I'm pretty sure the time step is much less than one year.
There is in fact an annual cycle in global *absolute* temperature (not temperature anomaly, which removes the annual cycle). But the temperature is greatest near aphelion (northern hemisphere summer), in spite of the much greater insolation at perihelion. This is attributed to the large thermal inertia of the southern hemisphere oceans, making the northern-hemisphere response dominant.]
Steve Bloom // April 7, 2008 at 8:26 pm |
Tom Woods, you’ll note that the same denialist sites that make (over-)much of short-term climate oscillations such as the PDO, cast derision on GHG physics and the models, and blow up the sun (as it were), spend very little time on deep-time paleoclimate, which IMHO is far and away the strongest argument for AGW theory. Looking at climate on a 10ky+ scale, the solar influences disappear except for the very long-term brightening and the Milankovitch cycles, leaving an incontrovertible association between GHG levels and global temperatures. The very long-term relationship is convoluted with plate tectonic and biospheric effects, complicating the explanation a bit, but in addition we have two very good examples of relatively short-term pulses of CO2 resulting in quite severe temperature excursions (the PETM and the Toarcian).
The scientific understanding of all of the above has advanced greatly in the last few years. See here for the big picture and here regarding the excursions.
A little OT, but while I’m on the subject:
The above tells us that CO2 levels of 450 ppm will melt the permanent ice given sufficient time. Other recent research about the response rate of ice sheets to warming tells us that it probably won’t take long. Jim Hansen has put two and two together here (and note that his co-authors on this include a representative grouping of leading paleoclimate experts).
David B. Benson // April 7, 2008 at 8:48 pm |
The time step depends upon which model on is talking about. A typical value for general circulation models these days, I am lead to believe, is three hours.
Leif Svalgaard // April 7, 2008 at 9:24 pm |
Benson: three hours! In that case the models should be able to tell us the story, and my original comment [of the annual variation being a much better test] stands. You simply run the model with and without the annual forcing and see what difference it makes, if any. I’m assuming that one can get intermediate output from the model, say once a month, rather than at the end after 400 years.
Gavin // April 7, 2008 at 9:37 pm |
more like 20 to 30 minutes. And even less for dynamical substeps for doing the momentum equations.
Hank Roberts // April 7, 2008 at 9:39 pm |
David, good link re excursions.
http://www.geolsoc.org.uk/gsl/geoscientist/features/page2617.html
Well worth a serious read, especially for those who aren’t familiar with this already.
Leif Svalgaard // April 7, 2008 at 11:05 pm |
Gavin, can you enlighten us here. Do the models handle the annual variation in TSI? that is: is this taken as a continuous input and do the models say what the effect of that input is? how strong the response is? how strong the damping is? etc.
Or don’t you know? [sorry, couldn't resist]
David B. Benson // April 7, 2008 at 11:24 pm |
The local regional air quality research group uses up to 30 minute time steps in MM5. The hallway conversation ended with the point that the numerics limit the time step to that maximum to obtain sufficiently high quality results.
Leif Svalgaard // April 8, 2008 at 12:02 am |
and at any given time and place the albedo is a variable that much be modeled too. Is there a cloud here? is this grid point over tropical ocean with its very low albedo, which in a few hours changes to high albedo because clouds developed?
Leif Svalgaard // April 8, 2008 at 11:13 pm |
Trying again:
Gavin, can you enlighten us here. Do the models handle the annual variation in TSI? that is: is this taken as a continuous input and do the models say what the effect of that input is? how strong the response is? how strong the damping is? etc.
Hank Roberts // April 8, 2008 at 11:33 pm |
Leif, amateur reader comment, just to help me understand the question — I found a whole lot of measurements available, and wonder which ones you’re asking abou using in models, and how:
http://www.atmos-chem-phys.net/7/3153/2007/acp-7-3153-2007.pdf
Contrails/clouds
http://www.springerlink.com/index/CN712037720740PT.pdf
The Middle Atmospheric Ozone Response to the 11-Year Solar Cycle
… eclipse of 29 March 2006 … surface ozone … solar ultraviolet radiation…– time resolution 30 seconds
http://www.atmos-chem-phys.net/8/425/2008/acp-8-425-2008.pdf
And I know the plankton biologists are modeling their species’ variation with such changes, to add to models!
Are you asking if the measured numbers from the existing satellite records of TSI are fed into the models Gavin works on?
Hank Roberts // April 8, 2008 at 11:51 pm |
Parenthetically, an observation from William Connolley in answer one of my questions about how losing the summer Arctic sea ice, for anyone who might know:
[I don't think albedo is very angle-dependent, at least not in climate models. In reailty, it might be different, and waer might be a special case (see Eric Swanson on sci.env) -W]
I was thinking of the glare from either water or ice, at a low sun angle, and wondering whether an ice-free Arctic in 24-hour sunlight would has any albedo calculated compared to an ice-covered summer Arctic with the same low sun angle but 24 hour illumination.
dhogaza // April 9, 2008 at 12:10 am |
Also, Leif, Gavin has only posted here extremely infrequently, so his silence probably is only a reflection of his not visiting the blog since.
You could always e-mail him …
Leif Svalgaard // April 9, 2008 at 2:21 am |
Hank asked: “Are you asking if the measured numbers from the existing satellite records of TSI are fed into the models Gavin works on?”
No, that would be too much to ask, because then he could only model 1978-2008, but the solar cycle and random variations of TSI are an order of magnitude smaller than the regular annual variation, which does not change over millennia, so he should use just the average well-know variation. No way to weasel out of that one because of “lack of data” :-)
Hank Roberts // April 9, 2008 at 4:01 am |
OK. But can you model TSI at the top of the atmosphere, or are you looking at components of SI “all the way down” to where it’s affecting plankton in the upper ocean? No well known average below the top of the atmosphere is there?
Julian Flood // April 9, 2008 at 7:42 am |
http://www.metoffice.gov.uk/research/hadleycentre/CR_data/Monthly/NMAT_SST_LSAT_plot.gif
and your SSN plot above would make a really nice comparison. Eyeballing, it looks like one goes down when the other goes up, with a lag of ten years, but that’s just a cursory glance. Is there an easy way of cleaning the SSTs of corrections/random volcanoes etc and producing the comparison in graphical form?
JF
Leif Svalgaard // April 9, 2008 at 11:21 pm |
Hank asked: ” But can you model TSI at the top of the atmosphere, or are you looking at components of SI “all the way down” to where it’s affecting plankton in the upper ocean?”
TSI at TOA is a known quantity with a noise component of the order of 1/100th of the regular smooth signal. What happens further down is the responsibility of the model. , no? It doesn’t make much sense to use a 20 min time step on parameterized large-scale spatial mean values, does it? Or is thus ignorance on my part? We have a similar situation in trying to model the solar dynamo where the time step is typically one day, but the input data is smoothed over six months.
TCO // April 10, 2008 at 12:24 am |
Leif: Keep pushing. Truth needs to be pursued whether it is SM using an AR(581) to model noise or warmers not showing the variability of solar input over time. “You go, girl.”
Hank Roberts // April 10, 2008 at 12:27 am |
> further down is the responsibility
> of the model. , no?
I’d have thought it was the responsibility of the field researchers first — to have collected some measurements all the way down through the atmosphere and upper ocean — and those would be eventually attractive to the modelers.
But I’m just a bystander ….
Hansen's Bulldog // April 10, 2008 at 12:30 am |
I don’t work on computer models of climate. But frankly it seems to me overwhelmingly likely that yes, computer models do include the variable solar insolation due to the eccentricity of earth’s orbit. It’s a straightforward and well-known phenomenon and a sizeable variation.
Hank Roberts // April 10, 2008 at 12:31 am |
Oh, a few cites:
http://lgmacweb.env.uea.ac.uk/green_ocean/workshops/workshop05.shtml
http://luv.dkrz.de/publications_2003/pub_47_104.pdf
GEOPHYSICAL RESEARCH LETTERS, DOI:10.1029/
Bio-optical feedbacks among phytoplankton, upper ocean
physics and sea-ice in a global model
I’d guess that the small measured TOA solar variation might well show up in biological cycles in plankton, much more clearly than it will show up in an overall total average Earth temperature number.
Hank Roberts // April 10, 2008 at 12:36 am |
Oh, and — more suggesting you might find the solar variation signal in the plankton work before you find it in the average global temperature:
http://www.springerlink.com/content/9846547065666123/
Abstract The influence of chlorophyll spatial patterns and variability on the tropical Pacific climate is investigated by using a fully coupled general circulation model (HadOPA) coupled to a state-of-the-art biogeochemical model (PISCES). The simulated chlorophyll concentrations can feedback onto the ocean by modifying the vertical distribution of radiant heating. This fully interactive biological-ocean-atmosphere experiment is compared to a reference experiment that uses a constant chlorophyll concentration (0.06 mg m−3). It is shown that introducing an interactive biology acts to warm the surface eastern equatorial Pacific by about 0.5°C. Two competing processes are involved in generating this warming: (a) a direct 1-D biological warming process in the top layers (0–30 m) resulting from strong chlorophyll concentrations in the upwelling region and enhanced by positive dynamical feedbacks (weaker trade winds, surface currents and upwelling) and (b) a 2-D meridional cooling process which brings cold off-equatorial anomalies from the subsurface into the equatorial mixed layer through the meridional cells. Sensitivity experiments show that the climatological horizontal structure of the chlorophyll field in the upper layers is crucial to maintain the eastern Pacific warming. Concerning the variability, introducing an interactive biology slightly reduces the strength of the seasonal cycle, with stronger SST warming and chlorophyll concentrations during the upwelling season. In addition, ENSO amplitude is slightly increased. Similar experiments performed with another coupled general circulation model (IPSL-CM4) exhibit the same behaviour as in HadOPA, hence showing the robustness of the results.
steven mosher // April 10, 2008 at 1:59 am |
I passed Lief some code for orbital parameters in ModelE.
To be EXACT, i found A routine, gave him the header, and a link if he wants to waste more time than I did. figuring out the simple ass question
he asked gavin.
Gavin should just give him a link to the code, scientist to scientist. answer the fricking question. MOVE ON.
Christ almighty
[Response: I find your comment, and your tone, very offensive. I don't know how often Gavin reads this blog, but I suspect it's not very often, and that the reason he didn't answer Leif's question is that he *didn't even know it had been asked*. And the right "scientist-to-scientist" way to get the answer would be to send an *email* to someone who works on computer models. Leif is a scientist, he certainly knows how to identify the modellers and get email addresses for 'em, and there are plenty to choose from.]
steven mosher // April 10, 2008 at 2:05 am |
Tammy, The routine is ORBPAR in modelE.
( I”M GUESSING!!!!)
See GISS modelE source browser. I’d be a idiot if I pretended to understand it. but I think it might answer Leif’s question. I sent him the link to source and the routine that I am GUESSING!!! addresses his concern. maybe gavins busy. but anyway. go to giss. modelE.
source browse. Find the routine ORBPAR
( orbital parameters) I think thats a good place to start.
If you have Qs, I can link to where I think the stuff is..or I should say SUPPOSE the stuff is
Hansen's Bulldog // April 10, 2008 at 2:10 am |
Leif, if you do find the answer you’re looking for I’d be grateful if you post it here.
Gavin // April 10, 2008 at 2:32 am |
My, my, what goes on when one steps out for a second….
The answer is yes and yes depending on what you actually mean. TSI at 1 AU varies every year according to the solar activity reconstruction you use (Lean et al for instance). The TOA solar input varies as a function of the time of year and latitude because of the orbital configuration – derived from the valid precession, obliquity and eccentricity for whatever period you want – the standard model uses values from 2000 AD with a vernal equinox fixed at Mar 21 hr 0. You can set this in the config file to be whatever you want from roughly a million years ago to a million years hence based on the calculations by Andre Berger. Note however that variations in the seasonal timing related to the fact the year is not 365 days is not taken into account (i.e. no leap years) and we haven’t set it up to adjust the precession every year. But those are small issues.
PS. Emailing me is far more efficient.
dhogaza // April 10, 2008 at 3:11 am |
So, Mosher, in your mind, NASA should change Gavin Schmidt’s job description by adding the phrase:
READ EVERY RANDOM QUESTION ADDRESSED TO YOU IN THE BLOGOSPHERE, AND ANSWER PROMPTLY.
Personally, as a tax payer, I’d prefer him to work at the job we pay him to do.
If he has time to stop by from time-to-time, great. If he doesn’t, STFU.
dhogaza // April 10, 2008 at 3:16 am |
Shit, I suggested that way back when, but that didn’t stop Mosher, did it?
Mosher,
The question wasn’t asked scientist-to-scientist in the first place. So MOVE TO REALITY.
Note: I don’t blame Leif in the least, he posts infrequently but apparently reads quite regularly , and perhaps thought Gavin does the same.
He gets a freebie on it.
But, Mosher, dude, you display exactly the kind of behavior I associate with CA (which is probably why you love the site so much, and participate so energetically).
That wasn’t a compliment, BTW.
[Response: Nothing to see here, folks ... move on.]
Leif Svalgaard // April 10, 2008 at 3:16 am |
Gavin and Steven M: thanks for the information. And since I think this was general interest the email route seemed less attractive. I’m against ascribing bad motives to people, so “no answer” did not imply anything nefarious to me. But my basic question remains: do the models show that the annual variation is dampened out? if so, there is a whole host of standard questions about dampened system that will tell us a lot about the sensitivity of the system.
jacob l // April 10, 2008 at 4:32 am |
Tamino did you check the differences between summer and winter anomalies?
like you I suspect that the earths climate would respond to changes in energy, from sun, atmosphere or other.
do you have any idea how long of a time series you would need to find the solar-cycle/ temperature connection?
thanks jacob
dhogaza // April 10, 2008 at 4:44 am |
I’m sure, if you’d asked in e-mail, that Gavin would agree to having his response made public here, if you’d asked.
Since you agree that there is nothing nefarious about Gavin’s not answering, and since you don’t jump into every thread here yourself (be careful of painting your own kettle back) …
Why don’t you simply ask him, via e-mail?
As he himself suggested above?
Sorry, Leif, you’re acting like an obstructionist, not one interested in honest dialogue, after Gavin himself said “geez, why don’t you just e-mail me?”
[Response: I don't see any evidence of misbehavior on Leif's part. There are enough situations which justify righteous indignation *without* seeing them around every corner.]
Adam // April 10, 2008 at 9:24 am |
You could email directly, then post the answer(s) on here (with permission)?
Like Hank did with WC’s answer above (okay he didn’t email, but it was a direct question on his blog which probably amounts to the same thing).
steven mosher // April 10, 2008 at 1:37 pm |
sorry about the tone, i assumed Gavin had read all of leifs questions when he answered the one about the time step. i was wrong.
Leif Svalgaard // April 10, 2008 at 2:12 pm |
now all we need is a similar ’sorry’ from dhogaza and some more restraint in the future.
Hank Roberts // April 10, 2008 at 3:47 pm |
And background music.
luminous beauty // April 10, 2008 at 3:54 pm |
From the world’s tiniest violin.
P. Lewis // April 10, 2008 at 4:01 pm |
Hank: Here Comes the Sun? House of the Rising Sun? The Sun Will Never Shine?
Take your pick.
steven mosher // April 10, 2008 at 5:05 pm |
I prefer this
http://www.youtube.com/watch?v=nKKpoCy0a5Y
always improves my mood.
[Response: I should delete this because it's just too far afield from the relevant; this is not a jazz blog. But it's too good not to allow.
But please, no more! This is about as off topic as it gets.]
ToddPT // April 13, 2008 at 4:33 am |
And astronomer has posted sunspot photos on another website, from April 12, which shows absolutely no sunspots, which he says is a first in 40 years of his observations.
He’s comparing it to the Maunder Minimum.
If these conditions indeed repeat, what predictions should we expect of the climate change theory?
A mini-ice age rather than a little-ice age? Can CO2’s forcing completely mask it?
[Response: Go tell that "astronomer" that telling lies is a sin. The absence of sunspots is not at all unusual during solar minimum.]
Bob // April 14, 2008 at 3:07 pm |
The 179 year cycle is well documented. It includes the Sporer, Oort, Maunder, Daulton minimums and others. This cycle would indicate we are approaching another minimum period.
Scientists that sudy the solar dynamo are predicting a very quiet cycle 25, which is also predicited by the study of solar intertial motion.
Lower temperatures (from proxy data) and solar minimums (also from proxy data) are correlated.
[Response: Where is that 179-year period "well documented"? Which scientists are predicting a very quiet cycle 25? Where is the correlation between lower temperature and solar minima shown?
Here's an opinion of mine: the claim regarding solar inertial motion is absolutely a crackpot theory.]
Bob // April 14, 2008 at 6:15 pm |
Thanks for the open scientific discussion.
Bob // April 14, 2008 at 6:34 pm |
Solar cylce 25 prediction by NASA
http://www.spacedaily.com/reports/Solar_Cycle_Appears_Headed_For_Historic_Low_Point.html
Correlation of solar cycle length and climate
http://www.tmgnow.com/repository/solar/lassen1.html
http://sesfoundation.org/dalton_minimum.pdf
Solar Inertial Motion
http://www.griffith.edu.au/conference/ics2007/pdf/ICS176.pdf
Hank Roberts // April 14, 2008 at 6:54 pm |
Bob, Google Scholar finds only two sources from which you might get this claim, both quite old:
http://scholar.google.com/scholar?sourceid=Mozilla-search&q=%22179-year+period%22+%2Bsolar
What’s your source?
Discussion? First we need to know facts behind the beliefs. What source?
Hank Roberts // April 14, 2008 at 6:58 pm |
Tamino, perhaps a clue — the one recent paper citing the first of those two old ones is indeed a statistical treatment of solar cycles; abstract here:
http://www.aanda.org/index.php?option=article&access=doi&doi=10.1051/0004-6361:20077574
Bob, have you read any of the full papers? Can you say something about them? Or what source you’re relying on?
Bob // April 14, 2008 at 7:09 pm |
referenced here as 178.7 years
http://www.ann-geophys.net/18/399/2000/angeo-18-399-2000.html
[Response: Stick around. I'll have a whole lot more to say about this in an upcoming post.]
Hank Roberts // April 14, 2008 at 7:35 pm |
More from solarcycle24, uh oh, statistics:
http://gltrs.grc.nasa.gov/Citations.aspx?id=330
Hank Roberts // April 14, 2008 at 7:37 pm |
Hmmm, looking for secondary sources:
http://www.google.com/search?aq=f&hl=en&safe=off&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&hs=MWU&q=ann-geophys.net+angeo-18-399-2000&btnG=Search
guthrie // April 14, 2008 at 7:47 pm |
Ah ha!
Thats the source of the wibble that a denialist was being coy about on my local national newspaper. They kept making comments about planets and the solar system and about someone having worked something out, etc etc.
It seems, on the surface to be plausible, although I have no doubt Tamino will dig the statistics up, however I predict that it will not in any real and large fashion affect the IPCC science.
Hank Roberts // April 14, 2008 at 8:18 pm |
9 citing Charvátová (some are comments, one self-cite; each of these has subsequent citations to follow up — but hasn’t anyone already done this?)
http://scholar.google.com/scholar?hl=en&lr=&safe=off&cites=17462717642088050626
Weak correlation: http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6TVM-4MYVG5S-1&_user=10&_origUdi=B6VHB-4B3JVDS-1&_fmt=high&_coverDate=06%2F11%2F2007&_rdoc=1&_orig=article&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=303bd3cee1b51a11fe079c7c1b7109d1
(Maybe this is why the Czech president has such a strong opinion …. nah)
Hank Roberts // April 14, 2008 at 8:21 pm |
Here’s the abstract from that latter cite, and the outline of the article, in case someone has access to the full text. Note the ranges of years in which they found a correlation.
Abstract
We study possible interrelations between the 300-year record of the yearly sunspot numbers and the solar inertial motion (SIM) using the recently developed technique of synchronization analysis. Phase synchronization of the sunspot cycle and the SIM is found and statistically confirmed in three epochs (1734–1790, 1855–1875 and 1907–1960) of the whole period 1700–2000. These results give quantitative support to the hypothesis that there is a weak interaction between the solar activity and the SIM.
Keywords: Sunspot cycle; Solar inertial motion; Phase synchronization; Hypothesis testing; Surrogate data
PACS classification codes: 05.45.Tp; 05.45.Xt; 95.75.Wx; 96.60.Qc
Article Outline
1. Introduction
2. Solar inertial motion and sunspots: The data
3. Synchronization analysis: The method
4. Results and their statistical evaluation
5. More statistical testing
6. Possible physical mechanisms underlying the synchronization
7. Conclusion
Acknowledgements
References
Bob // April 14, 2008 at 8:47 pm |
Hank – I quickly read through your first paper (I need to read it in detail when I get some quiet time). I’m sure we can fire off papers to each other all day that seem to contradict each other.
My comment about this specific paper is it did not reference the angular momentum of the sun and how it (the sun) “orbits” (for lack of a better term) around the barycenter of the solar system. This is really the crux of the solar intertial motion hypothesis.
Bob // April 14, 2008 at 10:32 pm |
Hank – Here is a better reference on the 179 year cycle. I was in a hurry this afternoon and grabbed the first that I saw.
http://bourabai.narod.ru/landscheidt/extrema.htm
I have read up on a lot of this – but only when I get a few minutes of quiet time. I guess I need to start cataloging what I have read – not only to send as reference, but to be sure I remember correctly what I have been reading.
Look forward to a civil discussion.
Hank Roberts // April 14, 2008 at 11:41 pm |
Bob, if you go to Google Scholar you can find the paper, or at least the abstract.
Sometimes that allows you to click somewhere in the margin to find subsequent work citing the paper.
http://www.springerlink.com/content/t775063131u00t84/
27 papers cite that one (including self-cites by the author in later work he published in “Energy and Environment” — Scholar doesn’t limit itself to refereed science journals, so you have to watch the sources even there)
http://scholar.google.com/scholar?as_q=LANDSCHEIDT+Solar+Physics+189+1999&num=100&btnG=Search+Scholar&as_epq=&as_oq=&as_eq=&as_occt=any&as_sauthors=&as_publication=&as_ylo=&as_yhi=&as_allsubj=all&hl=en&lr=&safe=off
Short answer seems to be the one I cited earlier — a weak correlation during three periods interrupted by long stretches without a correlation, whatever that means!
I’m not a solar physicist. I’m not a physicist at all. All I can say is “seems like not much interest in this and what’s published says it’s a weak association that doesn’t show up in climate measures.
Perhaps one of the real scientists will have more to say — but I’d suggest waiting til our host comes up with a topic, since he’s promised us one soon.
All I can say is — check cites, read citing papers, follow ideas forward, make sure you’re reading science journals, check authors’ other work.
“There might be a pony here somewhere.”
Hank Roberts // April 14, 2008 at 11:43 pm |
Oh, Bob, just to check, you wrote
> I quickly read through your first paper
Did you find and read the full text of it somewhere?
Bob // April 15, 2008 at 12:06 am |
Hank,
Yes – from your post at 7:35
Your link was to an abstract – but there is a link to the full PDF at the bottom of the web page. 34 pages in all.
Steve Bloom // April 15, 2008 at 1:01 am |
Re open scientific discussion, Bob, it has become rather obvious that with any sufficiently complex scientific subject, especially one involving lots and lots of apparent cycles, we start to see a lot of crackpottery. See e.g. Earl Happ’s contributions in the Svalgaard threads at CA.
Bob // April 15, 2008 at 1:48 am |
Steve, tried to find the thread without any luck.
I guess open discussion depends on who gets to define crackpottery?
Getting late – that’s all for today.
Bob North // April 15, 2008 at 2:48 am |
Steve – Thanks for the link to Oscar Peterson and HB thanks for leaving it in there. And it may not be as off-topic as you think. Weather and climate are probably more similar to jazz (quasi-cyclical interwoven riffs with lots of ad-libbing and apparently random variations on a theme) than to a Mozart concerto. Maybe modeler should take note.
BobN
Bob // April 15, 2008 at 12:34 pm |
Hank – thanks for the input Google Scholar is better than random web hits. I didn’t realize that some of the stuff on Springerlink could be accessed.
As usual, when I read through several of the papers I get conflicting results. But, I guess this is the way that science works. Papers come out, other try and duplicate results, some agree, some disagree and the entire process moves scientific understanding forward.
Bob // April 15, 2008 at 8:44 pm |
Hank – thanks for your discussion and helpful information. I was hoping to find a forum that could discuss some of these matters – but that seems pretty impossible here (not talking about you). Closed minds definitely prevail.
As I said in an earlier post – it all depends on who gets to define “crackpottery”.
No more from me – still looking for somewhere to discuss the various theories without being defined as an AGW alarmist or an AGW denier.
george // April 17, 2008 at 5:43 pm |
HB says:
Camp and Tung
There is something I don’t get about this.
This graphic from World radiation Center shows the deviation from min to max in the solar cycle close to 1 W/m^2.
If one assumes the standard factor (0.7) to account for albedo and the factor of 1/4 for the geometrical solar irradiance/climate forcing adjustment and further assumes that that 1 W/m^2 of climate forcing produces about 0.75 deg.C change in the global mean temperature (an assumption about climate sensitivity), that mans the min to max 1 W/m^2 would only produce about 0.13 deg C change in global mean temperature — only 72% of the change given by Camp and Tung.
And that would assume that global temperature responds pretty much instantaneously to the changes in solar irradiance (no time lag and no damping)
So, if we accept that the Camp and Tung result of 0.18C is correct, it would seem (I’m sure you will correct me if i am wrong :) ) that we have to assume that one or more of the other values we used — min to max irradiance change, 0.7 factor to account for albedo or 0.75C per 1W/m^2 climate sensitivity to forcing change — is not correct.
To get the 0.18C change that Camp and Tung claim, min to max change in solar irradiance would have to be about 1.4W/m^2 (all else being equal), but that would still assume that the climate system responded virtually instantaneously.
If one assumes that there is a time lag and that this produces damping the mismatch between the Camp and Tung value and the calculated is even worse, of course. It would seem that even a relatively short time lag of just a few years could make a significant difference (reduce the change in the global mean temp) given the fact that the solar cycle is so short.
Then again, perhaps the accepted value of albedo is not correct?
Or perhaps the factor of 0.75C for each W/m^2 climate forcing is not correct?
Or perhaps some combination of these?
Or perhaps I have missed something critical here?
[Response: The 1 W/m^2 variation throughout the solar cycle is very solid, as is the albedo 0.7. The climate sensitivity 0.75 deg.C/(W/m^2) is uncertain; a range of likely values is 0.5 to 1 (possibly higher).
Also, as I mention in the post, I doubt the result of Camp & Tung.]
J // April 17, 2008 at 7:24 pm |
The factor of 1/4 is also pretty certain, unless we’re mistaken about the shape of the earth.
:-)
David B. Benson // April 17, 2008 at 9:44 pm |
J // April 17, 2008 at 7:24 pm — Well, its actually a sort of pear shaped oblate spheroid, almost, so for very accurate work, 1/4 won’t do… :-)
Phil. // April 17, 2008 at 10:13 pm |
“Well, its actually a sort of pear shaped oblate spheroid, almost, so for very accurate work, 1/4 won’t do… ”
Actually it will, following Cauchy’s theorem any convex solid has a surface area/projected area of 4.
David B. Benson // April 17, 2008 at 10:44 pm |
Phil. // April 17, 2008 at 10:13 pm — I didn’t know that!
Learn something new every day. Thank you.
J // April 18, 2008 at 12:54 pm |
Hmmmm. Is that based on a mean projected area calculated over all possible directions, or something like that?
Because for a non-spherical solid, the projected area may vary depending on the projection vector, while the surface area is obviously constant.
I’m imagining an earth that’s shaped like a giant dime — the top and bottom are circles, while the side is infinitesimally thin. In that case, the surface area would approach 2(pi)(r^2), while the projected area would vary from ~0 to (pi)(r^2), depending on the viewing angle. If the mean projected area is 0.5(pi)(r^2), that would seem to fit nicely.
george // April 18, 2008 at 2:27 pm |
Strictly speaking, I believe that Cauchy’s “surface area/projected area” of 4 for convex solids applies to the case of an average of random projections.
But if you actually figure what the surface area/projected area is for the earth, using the surface area formula found here combined with the area of an ellipse and the relevant semi-major and semi-minor axis for the earth the answer differs almost not at all from 4 (I got 4.00005 when I plugged in the relevant numbers) Not enough (by a long shot) to boost the potential warming from 0.13C to Camp and Tung’s 0.18C, at any rate. :)
With regard to the Camp and Tung result (assuming the TSI variation and albedo are nailed down pretty well) suppose the sensitivity is actually near the upper part of the range — eg, suppose 1.1C per 1W forcing (instead of 0.75). That would be sufficient to make the calculated number for potential warming (0.19C) just (barely) enough to cover Camp and tung’s number (0.18C) .
But that would still be based on the assumption that virtually all (in that case 95%) the potential temperature increase from the solar output increase was realized over a very short time period (ie, that the time response to TSI increase was pretty much instantaneous).
george // April 18, 2008 at 3:00 pm |
With regard to the geometry issue I commented on above, the change in the ratio of “surface area/projected area” is actually in the wrong direction to help Camp and Tung, since we must divide the solar irradiance by a number (ever so slightly) bigger than 4, which would tend to reduce (not increase) the potential warming from the sun.
L Miller // April 18, 2008 at 3:14 pm |
I don’t think min – max change is what should be looked at, rather I think we should be looking at the amplitude of the sine wave which is 1/2 the peak – peak value. From this we should be able to approximate much above and below the mean the energy in the earth system varies due to the 11 year solar cycle.
A quick “back of the envelope” calculation seems to suggest the energy represented by the solar cycle is only sufficient to heat the atmosphere and top 100m of the ocean by +/- ~0.01 deg which is going to be pretty hard to pull out from the weather noise with only 11 years worth of data.
This isn’t to say there may not be some signal in the weather or that some feedback doesn’t come into play but as far as I can tell there isn’t enough energy involved with the solar cycle itself to make a meaningful change to the total energy in the atmosphere/ocean which drives climate. This doesn’t mean the forcing isn’t significant, it’s just not large enough to overcome the large damping (no pun intended) of the ocean over a mear 11 year cycle. Of course the other alternative is that my math could have been out to lunch. :o
Leif Svalgaard // April 18, 2008 at 3:40 pm |
george: “But that would still be based on the assumption that virtually all (in that case 95%) the potential temperature increase from the solar output increase was realized over a very short time period (ie, that the time response to TSI increase was pretty much instantaneous).”
Well, each year the TSI swings 90 W/m2 between January and July. Where is the response?
Phil. // April 18, 2008 at 4:01 pm |
Re J & George
Yes it is the average, I meant to type that but pressed send prematurely!
However when considering an average we have rotation E-W and ~46º rotation N-S so we do cover most of the possibilities. Even with the ‘dime’ example the mean comes out right (you can also calculate the variance). I published this approach for relating the measured equivalent sphere diameter to the sizes of regular crystals, it works well for compact shapes like cubes but less well for needle shapes for example.
george // April 18, 2008 at 5:54 pm |
L miller:
You may be right, but if so, it makes the Camp and Tung argument less plausible, since then even a sensitivity value of 1.1C for each 1W forcing would not account for the warming they claim. In fact, it would still be a factor of about 2 too low.
In that case, a change of 1/2 W/m^2 would have to account for the .18C change in global mean temperature, so sensitivity would have to be about 2.1C per 1W forcing. And that still assumes that all the potential change in the global mean temperature is realized over a very short time period.
Phil
Thanks for the clarification. Makes sense. I think the upshot is that “what ever way you look at it, the factor of 4 is pretty much correct for the case in question”. The impact of the potential slight deviation from “4″ in the case of the earth is going to be tiny compared to other uncertainty — in sensitivity, for example.
Do you agree?
Leif:
I can guess, but let me ask you outright: What is your take on the Camp and Tung result?
Paul // April 20, 2008 at 7:49 pm |
The connection is matching sunspot cycles to Accumulated Cyclone Energy. If you paste the latest cycle at the top, average winter temperatures in the middle and the ACE for each year at the bottom, then run a ruler across the page, you will get a chill up your back and down your legs. What is neat is the year 2001 where there is a saddle in the cycle, we had a worldwide drought followed with glacier activity in Glacier Bay Nat. Park.
I already filed my notarized research with NOAA, Hurricane Center, etc. and Colorado University Weather Team. Summary of first two years work, see “Track Sunspot Cycles”, Lakeland Ledger, 2 April
Preliminary study shows a connection in products. If this, then that. Boolean.
Carbons appear to be acting as a speedbump to Sunspot Activity. They absorb into the air when the air is warm and into the ocean when it is cold.
Study recession and growth marks in Glacier Bay Nat Park against your data.
Air temps depend on carbon stability. It’s like warming beans on an oven. The air above heats with it and circulates to much for a good reading.
Follow ACE, Glacier Activity and Carbon Dioxide Reflux.
Paul
Leif Svalgaard // April 20, 2008 at 9:43 pm |
george: I’ll quote what I’ve said over at CA [quoting what I've said here some time ago]:
RKeene // March 4, 2009 at 6:23 pm |
Could you do the same analysis, but on the integral of the sunspot number and the derivative of the temperature curve?