I really wish Gavin would put as much effort into getting the oddities with the GISTEMP dataset fixed rather than writing coffee table books and trying new models to show the sun has little impact.
This paper gets extra points for using the word “robust”. – Anthony
Solar trends and global warming (PDF here)
R. E. Benestad
Climate Division, Norwegian Meteorological Institute, Oslo, Norway
G. A. Schmidt
NASA Goddard Institute for Space Studies, New York, New York, USA
We use a suite of global climate model simulations for the 20th century to assess the contribution of solar forcing to the past trends in the global mean temperature. In particular, we examine how robust different published methodologies are at detecting and attributing solar-related climate change in the presence of intrinsic climate variability and multiple forcings.
We demonstrate that naive application of linear analytical methods such as regression gives nonrobust results. We also demonstrate that the methodologies used by Scafetta and West (2005, 2006a, 2006b, 2007, 2008) are not robust to
these same factors and that their error bars are significantly larger than reported. Our analysis shows that the most likely contribution from solar forcing a global warming is 7 ± 1% for the 20th century and is negligible for warming since 1980.
Received 17 December 2008; accepted 13 May 2009; published 21 July 2009.
Citation: Benestad, R. E., and G. A. Schmidt (2009), Solar trends and
global warming, J. Geophys. Res., 114, D14101,
doi:10.1029/2008JD011639.
hat tip to Leif Svalgaard
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Phil. (20:21:13) :
Bill Illis (18:47:20) :
Let’s recalculate gavin’s number (using his own assumptions) since he has to use a climate model to do his simple calculations.
Total solar irradiance change from 1900 to 2000 according to Lean 2004 – 2.0 watts/m^2.
Except as you know very well, because Lief has posted on here multiple times, Lean has revised that figure downwards. Of course that doesn’t suit your agenda so you keep going back to the repudiated data.
I was using the latest TSI data from Lean (2004 after she revised it downward) and the same data gavin posted up in his paper and that gives +0.158C from 1900 to 2000. Leif’s reconstruction produces a figure of about half of that.
Did I miss the bit where Gavin considers the energy budget effects of multidecadal changes in the rate of energy emission from the oceans and the constantly changing speed of the hydrological cycle ?
brazil84 (03:29:34) :
Slightly off-topic, but I have a quote I would like to nominate for quote of the week:
“When we apply full-blown GCMs, the match between hindcasts and observations is nothing short of impressive; ”
Yes – that is a peach. It just shows the self delusion these people suffer from. I honestly believe that some of them actually think that they are doing something worthwhile.
It’s like one of these huge financial institutions – none of which saw last years’s crash coming – turning round now and coming up with a model and saying that they have now hindcasted it. And that the results are ‘nothing short of impressive’.
John Finn: “almost constant” is not a fair description. The total forcing from the post-industrial rise in CO2 is ~5.35ln(387/280)=1.7W/m2. 15% of this is not negligible
But ~80% of that has occurred since 1940 since when we’ve had ~0.5 deg rise in temperature. You appear to be attributing a far smaller forcing in ~1940 to a warming which began 30 years earlier. What happened to the CO2 lag? i.e. the “heat in the pipeline” nonsense. Did the rules change around the middle of the century.
The CO2 forcing in 1915 was negligible when the warming began – and it wasn’t much more in the 1940s when the warming was over.
You also need to explain how the arctic warmed 4 times as much as anywhere else despite receiving far less sunlight (and none for 6 months). While you’re at it you can also explain how the Arctic cooled 4 times as much as anywhere else in the 1945-1975 period.
Simultaneously you seem not to believe that a lack of volcanic eruptions could cause warming, and that the lack of eruptions since Pinatubo is the cause of all the warming. Have a look at the data. You can see that there were no significant volcanic eruptions between 1912 and 1962. Clearly, a 50 year period with no major volcanic eruptions will be warmer than a 50 year period with three or four.
A significant warming trend began in ~1912 or just after. Of course there might be some short term warming after the effects of the stratospheric aerosols have cleared, but this is not going to last for 30 years. The effect of Pinatubo lasted about 2 years – or are you saying that 18 years later we’re still warming from that?
The 1915-1944 GISS trend was ~0.14 deg per decade. The trend was maintained throughout the period. It was not a short term uptick following a volcanic eruption.
That’s funny. I was under the impression that you did not accept this result. Please clarify
You appear to want me to clarify why your impression is wrong. I don’t know why your impression is wrong – it just is. Perhaps you didn’t read my post properly.
What a truly bizarre statement. Your position seems to be based on a determination to ignore all the well known, well characterised influences on climate, instead replacing them with a belief that no-one knows anything about why climate changes. Is that the case?
Spot on. The whole attribution issue is a crock and what’s more they know it over at RC. This is a response by Raypierre to a comment from one of the fan club.
[Response: Wayne, please note that this is Kyle’s article not mine, though I did encourage him to write it for us. I think the interesting question raised (though not definitively answered) by this line of work is the extent to which some of the pause in warming mid-century might have been more due to decadal ocean variability rather than aerosols than is commonly thought. If that is the case, then a pause or temporary reduction in warming rate could recur even if aerosols are unchanged. Learning how to detect and interpret such things is important, lest a temporary pause be confused with evidence for low climate sensitivity. –raypierre]
Oh sure there are the usual disclaimers in there, but they are finally being forced to accept that ocean fluctuations have a significant influence on climate. If ocean variability can pause and reverse warming then it can just as easily amplify warming. And that I’d suggest is pretty much what happened. There might be a CO2 signal but natural variability is the dominant factor.
I do like Ray’s “commonly thought” bit, though. I’m not sure what is definition of “commonly thought” is. It presumbly doesn’t include those they’ve been ignoring for the past 15 or 20 years.
RW (04:48:18) :
Allan M R MacRae: there is excellent CO2 data from before 1958 from ice cores. I used Law Dome data. Simply saying you think it’s “highly questionable” without any further elucidation is not useful.
So you like the ice core data.
How do you explain that it clearly suggests that CO2 lags temperature?
Please explain how the future causes the past.
*****************
As regards the UAH temperature data, when you fit a straight line through the warming portion of a sine curve, this will give you a misleading warming trend. The cooling portion preceded 1979, having occurred from about 1945 to 1975.
My bet is when the surface temperature data is analyzed and the warming bias removed, there will be no net warming from ~1940 to 2008.
*******************************
As regards cooling predictions, there are countless more since 2008, including a recent one (“no warming for 20 years”) from the warmist camp – but those are the easy ones, since Earth has been cooling for almost a decade.
Rather than defame those who had the courage to go against the conventional dogma of global warming pre-2008, perhaps you should consider how they came to their controversial conclusions.
There is an incredible amount of data supporting cyclical variation in climate and temperature – river flows, underwater sediments, and direct temperature measurements are a few examples.
***************************
The paradox is that you predict global warming , and I hope you are right – because humanity does much better during warm periods.
I predicted global cooling (in an article published in 2002, based on a conversation with paleoclimatologist Tim Patterson). I really hope that I am wrong, because humanity does poorly during cold periods.
*************************************
In any case, I am convinced that the body of evidence suggests that the sensitivity of Earth’s climate to a doubling of CO2 is much less than 1 degree C, and your side of this debate is promoting needless hysteria at great cost to society.
**********************************
http://www.wnd.com/index.php?fa=PAGE.view&pageId=104619
To translate:
Most Americans don’t believe AGW is a real problem.
Slightly off topic, but I found an interesting article in the online version of Esquire this morning. Its author, environmental writer Bjorn Lomborg, attacks Al Gore’s proposed solution to global warming and sugggests some much more practical–and less expensive–alternatives. Esquire isn’t Science magazine, but it seems to offer more evidence that at least SOME inroads are being made with the mainstream media. The article can be accessed here:
http://www.esquire.com/features/new-solutions-to-global-warming-0809
If Gavin’s 0.45C warming for each watt/m^2 of forcing is on the mark, the ~0.6C C20th warming (allowing for the negative PDO in 1900 Vs the positive PDO in 2003, and a bit of jiggery pokery with the temperature record) could be accounted for by solar alone if the ACRIM/Neptune measurements were used for estimating the TSI/sunspot ratio rather than PMOD.
Given that the sensors are not perfect, and degrade quite quickly and have to be allowed for with adjustments, it seems at least possible to me that small calibration errors could creep in at early stages in the program.
After all, the difference between Gavins assessment of solar contribution and it fully accounting for C20th warming is only 2.2W/m^2 or so out of 1366, when you take into account the location of oceanic energy absorbance that seems to be involved in the serious heat action on Earth.
An error of less than 0.3% in radiometer calibration caused by sensor degradation and imperfect data adjustment would cover that difference.
The stakes are high, and it’s no wonder the disgreement between the solar measurement teams is so ACRIMonious.
nogw’s link again to the Scafetta presentation to the EPA. The letters from the ACRIM team on around slide 16 make interesting reading.
http://yosemite.epa.gov/ee/epa/wkshp.nsf/vwpsw/84E74F1E59E2D3FE852574F100669688/$file/scafetta-epa-2009.pdf
tallbloke (23:37:59) :
According to Scafetta TSI was adjusted by 0.86 w/sq.mt. so it not longer follows solar cycle…but playstation games..
Gee, how does that play against the unadjusted GISS data?
Gavin, if you start with crap, you’re going to end with crap. In this case using GISS Temp data in your study was a bad starting point.
All the adjustments GISS does seems to push the more recent temps higher while lowering the older ones. Sorta defies the law of averages and runs against the UHI we now know has encroached on surface stations.
So you fault here was starting with a bad set of temperature data, after that point nothing else mattered.
Dave Wendt (22:49:17) :
Phil. (21:33:34)
“The OP referred to 2008! Currently the Arctic seaice extent is dropping fast, another 120,000 km^2 today, already passed 2008 and will probably be 2nd only to 2007 in a day or so. With the current rate of drift the NP webcam will be floating in the ocean in early Sept.”
The rather monotonous consistency of the drift patterns of the Arctic webcams over each of the years it has been deployed seems to me to strongly suggest that the primary driver of sea ice loss in the Arctic has not been in situ melting, but the relative strength of the Cross Polar Drift.
It’s both, the drift in both the Fram and Beaufort are major factors in the loss of multiyear ice.
Each of the sites that have persisted for up to a year has ended up in pretty much the same location at about lat 72-68 off the East coast of Greenland. The ice pinger data at this years site is showing only minor loss of thickness even though it has already moved to lat 85.
I think you’re misreading it, there’s 75cm of water on top over the last 20days, see the ice data below, note the isotherms.
http://www.arctic.noaa.gov/gallery_np_weatherdata.html
Allan M (03:38:02) :
Joel Shore:
“Steven Hill says:
How can 330ppm affect anything? It’s like a few grains of sand hidden in a gallon bucket of sand.
I am sure that we could find substances that would kill you at far lower concentrations.”
Phil:
“Nonsense, try reading up on the physics.
Add 330ppm of chromium to corundum (colorless) and you’ve got a ruby add iron instead and you’ve got a sapphire. Still think that a low concentration of an absorber can’t have a noticeable effect?”
—
Why do they have to give us these inane, fatuous, irrelevent analogies?
Because we keep getting posters making ‘inane, fatuous, irrelevent’ statements about the concentration of CO2 based apparently on their personal incredulity and lack of knowledge of the physics and chemistry
Helps to hide the non-sequiturs, I suppose. Anyway, when I studied it, the stuff Phil talks about was chemistry.
Well if the nonsense posts weren’t being made in the first place rebutting them wouldn’t be necessary. When I teach it it’s called Physical Chemistry, i.e. a mix of both.
———————
If I have it right we have a mixture of small model techniques, and GCM simulations in this paper. As I understand it most of the analysis is performed using small model techniques.
And it real seems to be a simple small model.
I have to confess that I like small linear models but they have to be fit for purpose.
In general the temperature function is produced taking forcings, applying them to the linear model and producing temperatures. Because the model is linear you can do this bt processing the individual forcings and summing the outputs to product the combined temperature function or sum the inputs (forcings) and produce an output temperature function directly. It makes no difference.
It is this linearity that allows linear regression to be used.
Now the big issue is what function should be use to map an individual forcing to temperature. In this case it seems that it is to multiply it by a constant and perhaps delay (lag) the result.
I am afraid that if you start (lagging) the result you are simply saying that the model is nopt fit for purpose. Forcings do not do time travel.
If you have a lag, you have to consider the attenuation of the signal associated with that lag.
What you need is a simple linear mopde that reproduces the lag and has the correct amount of attenuation. Now as we do not know the attenuation this is a subjective process. But there are siple linear models that are likley candidates.
The simplest mdoel to account for lags is the slab ocean model, this has its faults as it does not work well at all timescales, in particularly it can not cope with both long term trends and short term variations like volcanoes, ENSO, or the solar cycle.
A step up is the combined, thin slab and deep diffusive ocean model. This produces more acceptable results when considered accross a range of timeframes.
Now the model they seem to have used is the simplest of all the “no ocean” model, There is nothing in there model to produce either lags or estimates of the attenuation of the signal.
Now with the combined slab and diffusive ocean, you can produce significant amounts of attenuation with very little lag. So if you conduct an excercise using the data we are most sure about (the satellite data) you can find an attenuation of the signal by a factor of 1/2 quite easily, and still only show around a 1 year lag. So compared to the “no ocean” model your result would argue for a value twice as large for the significance of the solar effect.
Now I am sure (at least sincerly hope) that the authors know this. They say that such modelling is naive (the no ocean model is certainly naive) but there are simple linear models that do reproduce lag and estimate attenuation and they could easily use one.
A criticism of the combined slab and diffusive ocean is that it introduces two extra degrees of freedom, slab depth and a value for oceanic diffusivity. But they are constrained by OHC and seasonal temperature responses (lags and attenuations), so there should be no added degrees of freedom.
One thing I can say is that the “no ocean” model runs a significant risk of over emphasising the strnght of long period forcings such as GHGs at the expense of forcings that have significant short term variations, and if like solar, you have both a long term trend and significant short term variation this can lead to quite misleading results.
Alexander Harvey
Now this can be considerable.
@ur momisugly Tallbloke – The one that’s at the top of this thread. I guess it’s Gavins.
@ur momisugly Brazil84 – “Here there be skeptics”. 😀 . I like it. Good sugg. 🙂
Gee, how does that play against the unadjusted GISS data?
There isn’t any unadjusted GISS data. GISS starts out with NOAA fully adjusted adjusted NOAA data. It never encounters the unadjusted stuff.
evanmjones (07:00:36) :Read tallbloke (06:14:48) :
Good to see a scientific paper on Global Warming, Well Done Gavin.
I’ve always wondered how we have been at the low end of solar output in the cycle for the last 2 years, but somehow over the last two years temps have been going up, Remarkable.!.
I think that what we have here is a tacit acknowledgement that solar influence has a marked influence on climate. We cannot expect that the powers that be concede the point immediately.
As opposed to the IPPC stance of no discernible impact we are now getting recognition that solar cycles play a role.
What I find noteworthy is the .1 to .2 degree possible impact that has been quoted by Gavin using only TSI and ignoring other parameters such as the magnetic impact and the possible change to cloud cover.
Progress indeed. Maybe we will get to the truth sooner than we thought. Let us give credit where credit is due, if the likes of Gavin etc. concede that they may have been wrong and proceed accordingly with real and unbiased science to arrive at the true state of the issue whether it be CO2 forcing, solar, orbital or whatever then great.
Wishful thinking?
Bill Illis (05:22:42) :
I was using the latest TSI data from Lean (2004 after she revised it downward) and the same data gavin posted up in his paper and that gives +0.158C from 1900 to 2000. Leif’s reconstruction produces a figure of about half of that.
Lean doubts now that there is any long term trend. [Agreeing with me]. At the SORCE 2008 meeting in Santa Fe, she put it this way: “Long-term trends: do they exist?” This is science-speak for not trying to counterdict one’s earlier papers 🙂
“Our analysis shows that the most likely contribution from solar forcing a global warming is 7 ± 1% for the 20th century and is negligible for warming since 1980.”
I have not read all the comments, and I will not read this paper, so someone tell me what physical basis he uses to assert that the sun has variable input that was reduced post 1980. Why did the sun influence global temperature pre-1980, but not later. What, does increasing CO2 block incoming radiation? To me, this is unmitigated garbage, as Leif should agree (and already may have).
Jimmy Haigh says:
The folks at RealClimate are well aware of this “paper”: http://www.realclimate.org/wiki/index.php?title=G._Gerlich_and_R._D._Tscheuschner and those of us who are physicists are really, really embarrassed that such junk actually saw the light of day in a minor (but not completely disreputable) physics journal. Their claim that the greenhouse effect violates the 2nd Law of Thermodynamics is so easy to demonstrate incorrect that one could take students in a first year physics course through it on a problem set.
What climate models predict is the forced component of the climate’s response. They do not predict the exact course of internal climate variability (in essence, weather), which is very sensitive to initial conditions. There has been some recent work (by a group at Hadley and by Keenlyside et al) to try to initialize the ocean well enough to predict some of this variability over a timescale of a decade or so, which may be possible because the timescales for some processes in the ocean are slow enough that such a decade can still be short enough that the sensitivity to initial conditions hasn’t yet caused a complete divergence, but it is still very experimental at this stage.
If you want a good analogy, think of this: If we run a numerical weather model in January six months into the future, the weather prediction that it makes for a particular day will be garbage (and we will get a completely different prediction if we make very small changes to the initial conditions). However, it will correctly tend to predict the climatic consequences of the seasonal forcing, e.g., it will show that here in Rochester the climate in July is much warmer than in January.
What it is more logical it is the assesment by UN´s FAO study, that temperatures and fish catches follow LOD and PDO 60 years cycle, to which Scafetta refers as, at the end of his lecture, as “a symphony” of the solar system, following Kepler´s “music of the spheres”. Of course unacceptable for presumptuous “new age” scientists.
I’m having difficulty understanding how the demonstration that climate models don’t respond to solar forcing is useful in any way….
The Iceberg (07:33:45) : Nice cheerleader costume.
Abraham Lincoln think Michael Bay is uncool
http://lincolnsaysuncool.wordpress.com/2009/07/22/lincoln-says-uncool-michael-bay/