Click images for more details



Recent comments
Recent posts
Currently discussing

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

Powered by Squarespace

Discussion > Evidence, confidence and uncertainties


Busy, so this will have to be brief (which I regret, given the detail of your previous):

- I'm not looking for 'something nasty'. I do perceive a deeply-rooted bias.

- You ask me to demonstrate etc. Perhpas you might consider the existing body of work that supports a CS of ~3C to 2xCO2. Constantly trying to cast doubt by oblique means is failing here. I am not going to repeat my comment at Jan 17, 2012 at 7:49 PM, but I'm tempted, I have to admit.

- Tsionis et al. (2007) is a purely hypothetical exercise (although an interesting one). You are treating it as if it had some evidential basis.

- Rial. Why do I feel as if I am actually discussing this with Roger Pielke Snr? You seem to be very heavily influenced by RPS. Perhaps overly so.

- From SvD:

Our inferred uncertainty range for climate sensitivity, constrained by paleo-data, is 1.2–4.3C and thus almost identical to the IPCC estimate. When additionally accounting for potential structural uncertainties inferred from other models the upper limit increases by about 1C.

From AR4 WG1:

Since the TAR, the levels of scientific understanding and confidence in quantitative estimates of equilibrium climate sensitivity have increased substantially. Basing our assessment on a combination of several independent lines of evidence, as summarised in Box 10.2 Figures 1 and 2, including observed climate change and the strength of known feedbacks simulated in GCMs, we conclude that the global mean equilibrium warming for doubling CO2, or ‘equilibrium climate sensitivity’, is likely to lie in the range 2°C to 4.5°C, with a most likely value of about 3°C. Equilibrium climate sensitivity is very likely larger than 1.5°C.

So no real disagreement in SvD with the ~3C estimate.

- Zachos exhibits proper scientific caution:

Perhaps the most important developments [in understanding] concern the glacial history of Antarctica, and the scale and timing of climatic aberrations.In the case of the former, it is evident that ice sheets have been present on Antarctica for the last 40 My, and over much of that time have been extremely dynamic, implying a high degree of instability and/or sensitivity to forcing. As for the aberrations, their mere existence points toward the potential for highly nonlinear responses in climate to forcing, or the possibility of unexpected anomalies in forcing.

- Federov et al. (2006). Been through this before. I said:

The point about Cenozoic climate is that it represents a long cooling. The ice-free Early Pliocene with its perennial El Nino was a legacy climate state from a deep past of much warmer states. The elevated GAT was forced by low cloud albedo and high atmospheric water vapour content, both consequences of the perennial EN.

Once the Ma-scale cooling caused NH ice sheets to form around 3Ma ago, the ice-albedo feedback began to drive the pronounced glacial/interglacial oscillations. There is no need to elaborate on this essentially straightforward evolution of physical processes.

Holocene climate is at the bottom of a U curve in sensitivity variation. Push it hard enough left or right, and you get a snowball or hothouse Earth. Push it so just so far and you get a glacial or a Pliocene-like climate. Federov et al. suggests that the latter is a possible outcome from CO2 forcing on a centennial scale.

- Lindzen and Spencer: "At the risk of stating the obvious, their papers present the evidence for extremely low CS"

No, they don't. We've been through this before too, so another re-run is tiresome. Here is an incomplete list of replies in the literature to Lindzen starting with his 'infra-red iris' hypothesis (Lindzen et al. 2001) (ABSTRACTS ONLY):

Hartmann & Michelsen (2002)

Lin et al. (2002)

Harrison (2002)

Fu et al (2002)

Replies to Lindzen & Choi (2009)/Spencer & Braswell (2009):

Trenberth et al. (2010)

Lin et al. (2010)

Murphy et al. (2010)

Dessler (2010)

Replies to Lindzen & Choi (2011)/Spencer & Braswell (2011):

Dessler (2011)

Jan 18, 2012 at 12:37 PM | Unregistered CommenterBBD


You mentioned you've been reading up about variable CS, and I'm definitely in the market for anything interesting in that area. Do you have any links you can pass on?

I agree with you that it would be interesting to dig into the assumptions and methods behind the albedo measurements. I know there has been a lot of argument in the literature about the measurement of whole earth albedo, whether based on satellite measurements of clouds or on earth-shine.

Jan 18, 2012 at 2:17 PM | Unregistered CommenterPhilip


No need to hurry, I'd rather you took your time about it.

"I do perceive a deeply-rooted bias."

Only for good science, I promise you!

"Tsionis et al. (2007) is a purely hypothetical exercise"

Not quite sure what your point is here. Science is about making hypotheses, and confirming or rejecting them on the basis of evidence. These authors make a hypothesis and provide data and data analysis to confirm it. Looks like good science to me.

"Rial. Why do I feel as if I am actually discussing this with Roger Pielke Snr?"

Please, no more ad-homs! My best advice to you is to attack the arguments. You certainly know enough about this subject to do that very well indeed.

"So no real disagreement in SvD with the ~3C estimate."

I've no argument with that, or your comment re Zachos' caution, or your comment following Federov ... although a word of caution about your final comment on the U curve: the Z&G model is very crude and only designed to illustrate the general point. Do I read you right that you do accept my position regarding CS variation in the Holocene?

"No, they don't."

Oh, yes they do, although whether their evidence is right or not is a different question. I've certainly been over the S&B10 argument very carefully, including the maths, and I can find no obvious or egregious mistakes in it. The data they use may be incorrect or fraudulent, although I've not noticed any claims that that is the case, have you? Their data analysis uses a number of approximations, any of which may be invalid. However, it's basis is the same simple energy balance eqn used in Z&G and countless other papers and text books, which is widely accepted as being approximately valid. I've not noticed any specific criticism of the analysis they offer, only of their conclusions and interpretation - true certainly for Dessler, Trenberth and Barry Bickmore's blog article. On the other hand, SB10 contains very reasonable criticism of the analysis approach used in Dessler's reply and in other earlier papers. The only cogent criticism of SB10 that I'm aware of is the one I mentioned before i.e. they don't measure equilibrium climate sensitivity (whatever that is)!

Jan 18, 2012 at 2:27 PM | Unregistered CommenterPhilip


Sorry, busy day.

I can fly across the Tasman and see more white than blue. Over an entire day of varying solar incidence.

Tell me this has all been taken into consideration, carefully matriculated and fully represented in the Global Climate Models, someone? Like, exact difference in albedo that supports the quote above.

Yes, great care has gone into this. See here for an example.

Ice albedo has been subjected to equal scrutiny.

The generally used values are provided here together with a good deal of useful background info.

Hope this helps.

Jan 18, 2012 at 7:27 PM | Unregistered CommenterBBD


A bit of a cherry-pick on my part, but I'm just going to respond to this for now:

Do I read you right that you do accept my position regarding CS variation in the Holocene?

Holocene climate is at or near the bottom of the U curve of sensitivity. Here, where we are now climatologically, 2xCO2 is estimated to force a ~3C increase in GAT. AFAIK, there is no good evidence for significant variation in CS during the Holocene, although of course I may well be in for a surprise ;-)

Jan 18, 2012 at 9:06 PM | Unregistered CommenterBBD


Thank you for that, very helpful and much appreciated. I have been going through it and shall do some more reading.


I wish I could help but I think you are way ahead of me on that track. Also, I have been most unscholarly and neglected to note the sources I have been reading, but they include Hansen & Sato, Lindzen and BBD's favourite Science of Doom! There was also an interesting post on WUWT that tabled many T vs CO2 measures and arrived at a likely CS below AR4's LOWEST parameter. Let me try to find that for you.

Jan 18, 2012 at 10:09 PM | Unregistered CommenterGixxerboy


It looks like Jin et al are the 'go to' guys on this and their work on refractive measurement and modeling of rough versus flat water does seem to be thorough. Though the maths in the modeling is way above my pay grade :-)

However, one thing doesn't seem to have entered their calculations at all. I have contacted Thomas Charlock at NASA to see if he can shed any light on it.

Basically, their foundation work is done on a sea platform in Chesapeake Bay:

All very nice but, on reading through their work, they never contemplate any effect on albedo caused by sea bottom composition in shallow seas. They dismiss shallow waters early on as 'coastal' and insignificant. Now that may be the case if your world (ocean?) view is moulded by Chesapeake but neglects the fact that substantial areas of the world's oceans are shallow seas - I'm talking sub 12m. Overwhelmingly, these are in tropical, sub-tropical or equatorial latitudes. For example, The Great Barrier Reef and indeed most of coastal Australia including the Gulf of Carpentaria; the Arafura Sea; much of the South China Sea; the Southern Mediterranean, Red Sea and Persian Gulf; The Caribbean; The Java Sea etc.

Add these up (in addition to the entire world's coastal waters) and it's not an insignificant proportion. My point being, Jin et al seem only concerned with the optical effect of depth - up to 10m, 50m, 100 m etc etc. It's surely stating the obvious that the albedo above 8m of water on the grey clay seabed of the North Sea (or Chesapeake Bay, but their platform is in much deeper water) is significantly different to the white sand bottom of the tropics. (If I put glass in front of a grey slate it won't be much use as a shaving mirror. But silver?)

Given also that these tropical seas are exposed to a significant amount of the sun's effect, who knows how much difference this might make to the calculations?

Oh, and I couldn't see anywhere in Jin's papers consideration of whitecaps versus faceted water surfaces. Maybe Cox & Munk covered it. Probably not a biggie in the grand scheme of things.


Unfortunately WUWT was in 'black out' today in protest at SOPA/PIPA. Fair enough, but inconvenient. I'll dig out that CS article and post it.

Jan 19, 2012 at 6:02 AM | Unregistered CommenterGixxerboy

Oh, I should also say that the CERES satellite experiment didn't seem to have very good resolution. The Sahara shows up but the entire West Australian desert has the same albedo as Southern Victoria, Te Urewera National Park or the forests of the Pacific North West? Yeah, right...

Jan 19, 2012 at 6:10 AM | Unregistered CommenterGixxerboy


I'm such an oaf that I am struggling to find that study. But here's one from Princeton that gravitates around 2.4C:

Okay, I think this is it. There are only excerpts on WUWT and I cannot download the paper but it appears to be Gillet, N.P., et al., 2012.

That puts CS around 1.2-1.8 as far as I can discern. But that is conditional on transitory bases.

Give that a try and let me know what you think of it - I'm sure BBD will unleash hell's fire and brimstone on it!

Bottom line is that more studies see CS as constrained well below the IPCC AR4 top parameter and even possibly below the bottom. And the longer global T stasis endures (is it all going into undetectable OHC?) while CO2 steadily climbs, the more a lower CS looks plausible. (Could be ephemeral, of course.)

Good luck!

Jan 19, 2012 at 9:36 AM | Unregistered CommenterGixxerboy


First, very glad to hear you are asking researchers about their work directly. You will get far better info from Charlock about ocean depth/floor colour effects on albedo than from me (because I don't know). Tell me what he says.

I think the 2.4C paper is Schmittner et al. There are problems. No hellfire and brimstone, but if you want to see scientists talking about science (cf Richard Betts)...



And note Ray Pierrehumbert's comments:

- directly to Schmittner here:

- Schmittner co-author Nathan Urban links to his own new study which estimates CS @ ~3C here:

Planet 3.0 Urban interview here:

SkS here:

Gillett et al. (2012) investigates transient climate sensitivity rather than equilibrium CS. The general view is that TCS is ~2C and ECS is ~3C.

The estimates for TCS are dependent on the length of temperature time series used (1901 - 2000 and 1851 -2010) and subject to the greater uncertainty in temperature data pre-1900. Note that the 1901 - 2000 calculation yields a range of 1.7C - 2.5C - a good match with the widely-accepted best estimate from IPCC AR4 WG1.

IMO the lower and probably less reliable result has been over-hyped by 'sceptics'. With this in mind, also remember that it's one paper, based on a single model study. Even the authors feel compelled to warn:

We therefore recommend caution in interpreting the scaled projections derived from this single model, since our uncertainty estimates account only for possible errors in the magnitude of the simulated responses to the forcings, and not for possible errors in the observations, in the forcings, or in the spatio-temporal patterns of response to those forcings. We suggest that a similar analysis be carried out using multiple models once the necessary simulations are available, which will allow the effects of model uncertainty to be better accounted for.

Jan 19, 2012 at 2:35 PM | Unregistered CommenterBBD


One last thing:

Bottom line is that more studies see CS as constrained well below the IPCC AR4 top parameter and even possibly below the bottom. And the longer global T stasis endures (is it all going into undetectable OHC?) while CO2 steadily climbs, the more a lower CS looks plausible. (Could be ephemeral, of course.)

There is nothing - and I do mean nothing - in the literature that provides a convincing/robust estimate of CS much below 3C. I think the wise course now would be to reconcile oneself to a likely value of ~3C and begin to think about the probable implications.

As you probably recall, I was a 'lukewarmer' once too.

Jan 19, 2012 at 7:23 PM | Unregistered CommenterBBD


Thomas Charlock very kindly pointed me in the right direction and seabed albedo - and its variance - can be selected and run within the COART model. The assumed values equate wet sand to ' very dark sand', according to Tom C because sand darkens when wet. I could quibble about that because a lot of sand seabeds made from certain corals or shells remain startlingly white under water (I'm a diver amongst other pursuits). But it is unlikely to make a huge difference to a worldwide calculation. I haven't had a chance to fiddle with the values in COART and see what it gives. If you go to the CAVE page from your reference you can:

Hit “COART” on the right panel.
Scroll down to “Bottom albedo” and key in your desired value.
Hit “Submit Form” on the lower left.

Thanks for all those references. I shall have a look through.

I don't think the Gillett paper I referred Philip to is the one I had in mind, actually (sorry Philip). And, as you say, its focus is transitory sensitivity. I remember a chart from the other paper that showed the parameter 'cone' from AR4 and a CS track that was - at this stage - dropping below the lowest parameter. I shall have to bloody find it...or accept your ~3C suggestion. For now ;-)

Jan 19, 2012 at 10:03 PM | Unregistered CommenterGixxerboy


It's all right, my main interest is in the theory, so the arguments over the stats and data (i.e. the hard stuff) rather pass me by. It's possible we are still at cross purposes regarding CS, though. My expectation is that it is not a constant value, but that it varies over time. So although I'm interested to see the way that the different estimates have been made, my feeling is that they are all rather missing the point. This is why I tried to persuade BBD to write up for us how he personally understands the definition of ECS. I still hope he'll do that - it could provoke another interesting discussion!

Jan 19, 2012 at 10:42 PM | Unregistered CommenterPhilip


Yes, I'd like BBD to summarise what really convinces him of an ECS of ~3C: not just the numerous papers that support it, some (not all by any means) of which smack of dubious groupthink.

I think you are in good and crowded company in thinking that CS varies according to different prevailing states. Whether this is defined as TCS and ECS...? Perhaps.

IIRC, the study I mentioned focused on more modern observations rather than the usual paleo suspects to show that CS does not appear to be following IPCC guidelines. Properly calibrated instrumental records and observations beat paleo proxies and models hands down, in my book. the timesacales involved of course mean this can be categorised as TCS.

Good luck with your study. I shall find some time to address BBD's copious reading list (!) and I'll post anything that may interest you.

Jan 20, 2012 at 12:01 AM | Unregistered CommenterGixxerboy


Thanks for the good wishes, and good luck to you too with the list! If you want some light (!) reading in between, you could glance at this research group's site, if you haven't already - this is where I plan to spend some of my study time. They've been investigating natural processes in the climate for more than 30 years, with a great deal of success. They are not mentioned once in AR4! If nothing else, and you've not seen it already, this opinion piece highlights some of their ideas and concerns. All very stimulating - and amply confirms what a great scientific topic climate can make, with something for everyone!

Jan 20, 2012 at 12:01 PM | Unregistered CommenterPhilip


Yes, I'd like BBD to summarise what really convinces him of an ECS of ~3C: not just the numerous papers that support it

Jaysus! These papers are why I am convinced that the estimate for ECS is about right. Some of what you ask for is here, on this thread! I do not pretend to be a source of original wisdom on this topic. I leave that breathtaking hubris to other amateur climatologists and 'sceptics'.

Jan 20, 2012 at 6:20 PM | Unregistered CommenterBBD


This is why I tried to persuade BBD to write up for us how he personally understands the definition of ECS.

Apologies - missed this. In the usual way:

dT = λ*df

Where 'dT' is the change in the Earth's average surface temperature, 'λ' is the climate sensitivity (°C/[W m^2]), and 'dF' is the radiative forcing.

In Hansen & Sato 2011:

Climate sensitivity (S) is the equilibrium global surface temperature change (ΔTeq) in response to a specified unit forcing after the planet has come back to energy balance,

S = ΔTeq/F, (1)

i.e., climate sensitivity is the eventual (equilibrium) global temperature change per unit forcing.

Jan 20, 2012 at 6:23 PM | Unregistered CommenterBBD

1/ dT = λ*df is a linearization of the energy balance eqn, and so this definition is subject to the issues raised in Z&G that we discussed above.
2/ According to the H&S definition, the system must be in equilibrium not only after the perturbation has taken effect, but also immediately prior to the perturbation, and neither of these conditions is realizable in the real world.

In order for the definition of ECS to be measurable, it is therefore also necessary to specify how these factors should be accounted for. The difficulty in achieving this is presumably one of the reasons why it is so difficult to tie down the value in a robust manner.

Jan 20, 2012 at 7:40 PM | Unregistered CommenterPhilip

Consider a glacial or similar. It is quite reasonable to assume that the variation of T with F follows a curve with similar characteristics to the one shown in Z&G. In this situation, the ΔT = λ*ΔF definition of ECS corresponds to the slope obtained by drawing a line from a start point on the bottom branch to a finish point on the top branch, which is obviously very different to that obtained by drawing a tangent to the curve itself (which corresponds to the dT/dF definition I suggested above). Even in the modern world, a transition similar to those discussed by Tsonis and co would presumably have a similar (although smaller) impact on any estimate that uses a sufficiently long timescale.

I'm not trying to cast doubt on the work of the many scientists who have tried to estimate ECS and who no doubt use a variety of means to try to eliminate such problems. Instead I am simply trying again to explain why I doubt the justification for your insistence that ECS is with certainty ~3K. On the basis of the papers I've seen, dT/dF could easily turn out to be 1K or less or it could easily turn out to be 6K. For example, if you couple the IPCC attribution logic with the latest TSI reconstructions, then you can easily end up with an estimate of ECS of at least 6K! The failure to adequately address issues like this, is one reason why I personally have lost trust in climate science (in any case the IPCC version of it). This snip (again from Lovejoy and his colleagues) expresses a route that would certainly help to reestablish my trust at least:-

Nonlinear ideas have shown how to tame fractal and other nonclassical monsters, and these are important successes. Yet in the absence of societal support for very promising alternative nonlinear approaches, applications will continue to be deprived of this knowledge and resources will continue to be squandered on state of the art techniques informed by inappropriate theories. Thus, funding agencies, academic institutions, journal editors, and individual researchers need to see the future potential of nonlinear geophysics to solve science problems that have consistently been beyond the reach of traditional methods.

Jan 20, 2012 at 9:06 PM | Unregistered CommenterPhilip


Let's consider the evidence. There is clear convergence, by different methods, on an estimate of ECS for 2xCO2 of ~3C.

Why the convergence? I have a hard time believing that everybody is wrong. Too many different methodologies in play for that to be the case. And ~3C fits with paleoclimate better than lower values. It's enough for me, so that's where I hang my hat ;-)

Jan 21, 2012 at 12:54 AM | Unregistered CommenterBBD

BBD You may very well be right about the ECS value, but using the argument I have a hard time believing that everybody is wrong is unconvincing. There have been many examples in the past in science where everybody has indeed been wrong, even when different methods have been used.

Jan 21, 2012 at 5:55 PM | Unregistered CommenterArthur Dent


I agree with Arthur, you've not really provided a response to my arguments.

I understand that you like SoD, which is good if correct - so do I. I just noticed some recent articles relevant to the issues we've been discussing, starting here (it's a series of three so far).

Do these give cause to be a little more critical about the climate sensitivity issue? My impression from these articles is that (a) SoD is far more respectful about Roy Spencer than you, and (b) he is open to the idea that CS varies, both over time and space, and that this also would have implications for its measurement.

Jan 22, 2012 at 10:38 AM | Unregistered CommenterPhilip


I've summed it up at Jan 21, 2012 at 12:54 AM. I base my view on reasoned deduction from the available evidence. Not prior commitment to a low CS as is clearly the case with you.

AD doesn't say anything - just tries to inject doubt. He doesn't even bother to provide any examples.

Jan 22, 2012 at 6:49 PM | Unregistered CommenterBBD