Click images for more details



Recent comments
Recent posts
Currently discussing

A few sites I've stumbled across recently....

Powered by Squarespace

Discussion > Richard Betts: 20 C Temperature Attribution

Hi Richard,

In the Santer post, you commented about the attribution of temperature variations during the first half of the 20th C. : Sep 5, 2011 at 3:17 PM

The reason that the "natural only" warming is greater in the early 20th Century than later is because the increase in solar irradiance is greater then (early 20th Century).

You cited Lean in support of this. However, more recent studies than Lean's suggest that TSI remained approximately constant during early 20th C. For example, slide 16 in Please check the diagram: the difference is very significant and if correct does mean that the attribution you suggested is likely not correct.

Just to spell out the consequences ... If changes in TSI are not able to explain the temperature changes during the early 20th C, then a currently unknown factor was responsible. Late 20th century temperature change can no longer reliably be attributed to rising CO2 levels, because the same unknown factor could be responsible. Similar remarks presumably apply to attribution of ancient changes as well.

Just to be clear, I certainly accept the basics of climate physics and would therefore expect to see a rise in T since 1950 due to CO2 changes. I also strongly support the need to decarbonize energy supplies. Nonetheless, the attribution of early 20th C changes appears to be a mystery that needs and deserves a good answer.

Sep 7, 2011 at 6:11 AM | Unregistered CommenterPhilip

Hi Philip

Thanks, that's very interesting (and nice to have a scientific question!) :-)

I'll get back to you.

Sep 8, 2011 at 9:31 AM | Unregistered CommenterRichard Betts

Hi Richard,

Thank you very much - I'm curious to find out what the answer is.

Sep 10, 2011 at 5:33 PM | Unregistered CommenterPhilip

By coincidence, some more about this issue in section 4 of this paper from Bulletin of the American Meteorological Society:

The substantial warming during the period 1910-1940 has been attributed by nearly all the modeling groups to some combination of increasing solar irradiance and a lack of major volcanic activity. The cooling and leveling off of average global temperatures during the 1950’s and 1960’s is attributed primarily to aerosols from fossil fuels and other sources, when the greenhouse warming was overwhelmed by aerosol cooling.

Based upon new and more reliable solar reconstructions, the AR4 (Section concluded that the increase in solar forcing during the period 1900-1980 used in the AR3 reconstructions is questionable and the direct radiative forcing due to increase in solar irradiance is reduced substantially from the AR3. However, consideration of Table S9.1 in the AR4 shows that each climate model used outdated solar forcing (from the AR3) that was assessed to substantially overestimate the magnitude of the trend in solar forcing prior to 1980.

Sep 13, 2011 at 10:19 AM | Unregistered CommenterPhilip

Yes, there are many different 'reconstructions' of solar irradiance and they give very different answers. See for example IPCC AR4 figure 2.10 which shows that the numbers are all over the place. Most climate scientists cherry-pick the Lean paper, because it shows what they want - a rise in the early 20thC and then a leveling off.

See this post by Bob Tisdale for a good discussion of this issue.

It's important to appreciate that we only really know anything about TSI since about 1980.
Before that it is all 'reconstructions', based on speculative assumptions, which explains why different papers get very different answers (to be fair, the IPCC acknowledges that scientific understanding of this is 'low'). Note that the third author on the Lean paper is Ray Bradley, of hockey stick infamy, which sets alarm bells ringing. I would put more weight on papers written by solar physicists only, who have no axes to grind about climate implications.

Regarding the writing of IPCC AR4, see if you can guess which solar physicist the IPCC invited to be a lead author of chapter 2.

Sep 13, 2011 at 1:13 PM | Unregistered CommenterPaulM

Thank you, Paul. I can certainly see "Judith Lean" in the list of lead authors here. FWIW, both this and the Curry snip I gave above, suggests to me that there may be over-compartmentalization within the IPCC process.

Sep 13, 2011 at 4:26 PM | Unregistered CommenterPhilip

Hi Richard,

I'm sorry to keep on at you, but I'm think it's important to have an answer on this one, so I'm going to spell out my concern again:

You told the folks here that:

The reason that the "natural only" warming is greater in the early 20th Century than later is because the increase in solar irradiance is greater then (early 20th Century).

The attribution studies in AR4 are partially grounded in this observation, which derives I think from a 2000 study by Judith Lean. However later studies suggest that TSI was approximately constant during the period. Therefore, the validity of the AR4 attribution studies is called into question.

Apparently, the TSI and other forcings used in the attribution studies derive from data described in chapter 2 of WG1,

Amongst the lead authors of this chapter are both Lean and yourself. The most likely explanation I can see for your comment, is that there is inadequate communication between IPCC lead authors. I have to say though that this conclusion scarcely fills me with confidence about the report's contents.

Having said all this, I imagine I've simply got hold of the wrong end of the stick as usual. But if I have, why not tell me how?

Sep 18, 2011 at 2:12 PM | Unregistered CommenterPhilip

Hi Philip

Sorry to take so long to respond, I've been looking around in the literature. Has the Svalgaard curve been published - I've not been able to find it yet. Bob Tisdale's WUWT post didn't give a reference to a paper.

Oct 2, 2011 at 12:35 AM | Unregistered CommenterRichard Betts

Hi Richard,

I'm away from my laptop at the moment so can't double check, but from memory slide 16 in Svalgaard's presentation references the original sources:

Oct 6, 2011 at 11:20 AM | Unregistered CommenterPhilip

Hi Philip

Unfortunately not - there is a curve labelled "Svalgaard" but no reference to an actual paper. I've looked through his home page and on Web Of Science and google, and not yet managed to find any actual papers with that curve in. So when you're back online, it would be great if you could point me to any that I've missed.



Oct 6, 2011 at 12:14 PM | Unregistered CommenterRichard Betts

Hi Richard,

I think the slide I mentioned refers to similar results from a number of other authors, not only Svalgaard. The Curry paper I linked above may also contain the references you are after. She does reference AR4 - that may possibly contain the references. You could try here:

Failing that I'll find them for you when I get back next week.


Oct 11, 2011 at 6:11 PM | Unregistered CommenterPhilip

Thanks Philip. Yes, our AR4 chapter does cite some of the papers, such as Wang et al (2005) (which includes Judith Lean, and which I think is the "Wang" curve on the powerpoint slide although this is not actually stated), but it's the Svalgaard one I'm particularly interested in as it is the most different from the original Lean. Our AR4 chapter cites a Svalgaard paper on his geomagnetic index, and I suspect he's used that as a proxy to reconstruct the TSI in the powerpoint figure, but I can't find a paper either confirming this or showing how it was done.

If you have any better luck than me then please let me know, as this is quite interesting.

Oct 12, 2011 at 9:51 PM | Unregistered CommenterRichard Betts


The Svalgaard dataset is not presented in a journal article, but I understand from the author that it is based on a regression between the sunspot number and the observed TSI during the satellite era but without adding in any background to the solar minimum. The following links from his research page provide some more information about how the data was derived:,%20EUV,%20TSI.pdf

All of the data sets from Svalgaard's presentation are listed here:

The first graph in this piece shows that the "Preminger 2010" reconstruction is very close to Svalgaard's. This reconstruction is described in the following 2006 paper:

The "Krivova 2007" data is also close to Svalgaard's and is discussed here:

and also in this subsequent paper:

Oct 16, 2011 at 10:19 AM | Unregistered CommenterPhilip

Thanks Philip. I'll take a look!

Wonder why he didn't publish it in a journal yet though? Has he tried and been rejected, or just not got round to it yet? If you know the author it would be interesting to know.

Oct 16, 2011 at 8:26 PM | Unregistered CommenterRichard Betts

This is interesting, but the Lean reconstruction is from 2000. Everything more recent (that Svalgaard shows) is in better agreement with Svalgaard's unpublished result (although it is still the outlier 1700 - 1900 compared to the reconstructions by Wang and by Krivova).

I know exactly nowt about this field, but assuming that the major studies are those shown, Svalgaard seems in better accord with what is apparently current thinking.

As they say, 'following'.

Oct 16, 2011 at 9:05 PM | Unregistered CommenterBBD


I'm afraid I don't know the author, although he was kind enough to confirm one of the references I mentioned.

There is also a long discussion at Climate Audit which might throw up some further answers:

Oct 17, 2011 at 10:57 AM | Unregistered CommenterPhilip

Thanks Philip (and BBD - good to see you are following this!)

You may be interested to know that in our latest simulations for IPCC AR5 we are using one of the more recent TSI reconstructions which show the much smaller increase over the 20th Century - see figure 13 of this paper by by colleagues.

The TSI reconstruction is from Lean 2009 which uses Wang et al (2005) as cited in the Svalgaard ppt presentation and our IPCC AR4 radiative forcing chapter.

The impacts of this on attribution of the early 20th Century warming haven't been looked at yet (but will be). Looking back over our conversation back to your original question, maybe I've been worrying too much about trying to find the Svalgaard curve in a peer-reviewed paper yet - if the use of Wang et al means that solar forcing cannot explain the temperature rise then the even flatter reconstruction of Svalgaard may not matter anyway!

As you say, there are important uncertainties here. Will be interesting to see the outcome of D&A studies - I'll keep you posted.

Thanks very much for higlighting this very interesting topic!

Oct 17, 2011 at 4:02 PM | Unregistered CommenterRichard Betts


Wouldn't miss it for the world.

I do wonder about the implications for estimates of GHG forcing over the last century. Contrarians will point to the use of outdated TSI estimates to force the models referenced by AR4 WG1 and claim that it 'proves' GHG forcing is over-estimated. It seems to me equally likely that the opposite could be true.

Oh well, I suppose the whole report will have to be done again. What? AR5, you say? Oh...


Oct 17, 2011 at 4:36 PM | Unregistered CommenterBBD


Thanks for the reference to the paper by your colleagues, very informative. As you say, it will be interesting to see whether the upcoming model runs are able to exhibit at least the kind of changes seen in 1900-50. Unfortunately, the difficulties in reliably reconstructing forcings prior to 1980 are unlikely to go away, so I imagine that even in AR5 model based attribution will remain vulnerable to criticism of this sort.

Another interesting possibility is that the 1900-50 variations did not result in any straightforward way from external forcings such as TSI. This might be the case I think if there is substance to ideas discussed by Rial, Koutsoyiannis and others. One plausible suggestion is that the real world climate exhibits variations on all scales. I will be interested to find out whether the AR5 runs can also exhibit such behaviour.

I imagine you're well aware of these ideas, especially given that (as I understand) you got your Physics at Bristol. For anyone else reading this not quite so sure, I noticed the following quite nice informal articles:

By coincidence (again!) there is a highly relevant post from today's Climate Etc:

BTW, let me repeat that I don't believe that problems of model-based attribution should change the general conclusion of the need for CO2 emission reductions. I'm just rather interested to understand how far it is reasonable to hope models can go in predicting the future climate system.

Oct 17, 2011 at 6:54 PM | Unregistered CommenterPhilip


Thanks for your comments here. I totally agree with you, it is easy for either camp to take these results as support for their position. Plus ca change etc. I think it is important that the model runs use the best available data as input, and from what Richard says it sounds as though this is being done in the upcoming runs for AR5.

Oct 17, 2011 at 6:57 PM | Unregistered CommenterPhilip


Hello again.

I'm just rather interested to understand how far it is reasonable to hope models can go in predicting the future climate system.

I'll leave that to RB and perhaps Tamsin, if she can be persuaded to visit a civilised backwater like this thread.

Oct 17, 2011 at 7:17 PM | Unregistered CommenterBBD

Hi Philip

Yes I did do physics at Bristol. I take it you've been following me on Twitter then? (If so, thank you for your interest!)

As it happens, in my first year we were lucky enough to have a guest seminar on fractals by Benoit Mandelbrot. He had to apologise in advance to the over-capacity audience for the predominance of equations over pretty pictures!

Oct 17, 2011 at 7:54 PM | Unregistered CommenterRichard Betts


In fact I got it from your MO page (just checking up on people's credentials like a sensible boy!). I'm also acutely aware that you're not able to return the compliment - partly because my own background is not so relevant (I'm simply another anxious citizen) and also because at a personal level I've felt intimidated on some sites (although never here). FWIW you can if you want find a little more about my dim and distant science past here ('82), which should also explain why your Bristol affiliation caught my eye.

Oct 18, 2011 at 9:58 AM | Unregistered CommenterPhilip

Philip, if you did a PhD with Michael Berry I don't think you need to feel intimidated about your background!

Regarding future predictions, IPCC AR4 made some rather extravagant claims about this, with some dubious analogies involving coins and dice (see Ch 1 page 105 or here). Roger Pielke got a bit cross about this, describing the claims as absurd and scientifically unsupported. I wonder what MVB would think.

But this is getting a bit off-topic. Has anyone emailed Svalgaard to ask him whether there is a paper somewhere that you haven't found, or if there is one in the pipeline, or if not why not? (I think part of the reason is that he has had some bad experiences with peer review).

Oct 18, 2011 at 10:57 AM | Unregistered CommenterPaul Matthews


Thanks. I e-mailed Svalgaard asking for references for the "Svalgaard 2007" and "Preminger 2010" curves, with the outcome detailed in my comment @ Oct 16, 2011 at 10:19 AM. Regarding what MVB would think about this issue, I don't know. I did consider asking him, but decided that since I hadn't kept in contact it would be unfair to ask.

BTW, I clicked through to your google sites page - it is very useful.

Oct 18, 2011 at 12:17 PM | Unregistered CommenterPhilip