Okay, so 1C - 1.5C? Or below 1C?

Does it matter that much?

matthu

That you even ask demonstrates how little you actually understand.

Okay.
Philip says he's not sure of the point you are trying to make, and I'm definitely not sure of the point you are trying to make.

But you carry on being evasive and giving off supercilious airs - because I don't expect anyone cares and I don't think your questions matter.

Oh FFS matthu. Just go away and let this discussion proceed in peace.

BBD,

Sorry for the delay in getting back to you again. Please let me try and explain my position in another way.

Refer to the Zaliapin and Ghil paper we talked about before. Consider their equation 11, the standard 0-D energy balance model. They analyze this by substituting approximate formulas for the greenhouse effect and for the ice-albedo effect, and derive the graph of temperature T vs forcing F shown in figure 5. Sensitivity S is proportional to the slope of the graph (dT/dF), which as you can see is a function of temperature in this model. For small changes in forcing, temperature changes slowly except near the turning points. Furthermore, the existence of the turning points means that there are three possible temperatures for some forcings, and an associated hysteresis effect if a turning point is traversed.

The Z&G model only includes the ice-albedo feedback. In reality, there are many interacting feedbacks, and many variables of interest aside from temperature. The result is likely to be further foldings and wrinkles in the graph. The sensitivity is not a simple linear function, except for small forcings and over small timescales. The Schaffer review article I mentioned to you a while back, qualitatively discusses climate from the point of view of dynamical system theory and makes similar conclusions. So again, it may be that for small timescales, the Spencer and Lindzen calculations are right, but for longer timescales, there is the possibility of very speedy changes across turning points. See also this sobering paper, if you haven't already read it.

I think it's also interesting to contemplate the difference between the Pliocene and Pleistocene temperature reconstructions from this p.o.v. (e.g. HS11, figure 1b). If the data is reliable, then the Pliocene is more stable even though the continental arrangements are similar. Perhaps this is related to the proximity of major turning points in the temperature range experienced during the Pleistocene?

Philip

As you know, I find all this to be verging on avoidance, or misdirection. But I would be interested in Richard's take.

The sensitivity is not a simple linear function, except for small forcings and over small timescales.

How do you explain the relationship between CO2 ppmv and Cenozioc cooling?

Are you serious? Please try reading up a little on this topic, and then come to me.

Look, I'm also interested in your take regarding these issues, which is I why I took the trouble to have another go at talking to you about them. In many ways, the non-linear dynamical aspect of climate should be a cornerstone of climate physics, to be taken at least as seriously as radiative physics. It is certainly at least as respectable as a piece of theoretical physics, and utterly relevant to climate science. Why not at least take a look at the Thompson and Sieber paper I referenced if nothing else, I can't see how anyone could take this one as a "skeptical" paper (whatever that means).

Skeptical paper (n)

scientific paper journal editors may feel compelled to resign over should it ever be published as a result of exclusion-failure by cosy pal-review process

Regarding Cenozoic cooling, this paper from 10 years ago presents the view that there were multiple causes, including the "potential for highly nonlinear responses in climate to forcing" (although possibly you can cite some more recent updates?). In any case, my speculation was not related to the reasons for cooling during Cenozoic, but simply to the idea that the climate prior to the ice-age cycles appears from reconstructions to be have more stable than subsequently. If you take the reconstructions seriously, then I hope you'll agree at least that an explanation for this behaviour is required. Non-linear effects do potentially provide such an explanation (for example, look for discussion of LDR in the T&S paper). How would CO2 explain it?

Philip

If you take the reconstructions seriously, then I hope you'll agree at least that an explanation for this behaviour is required.

Yes, of course. Ice albedo feedback. Is that it? Is that your point - that climate on a world with ice caps is less stable than on an ice-free world?

This doesn't affect estimates of CS such as those undertaken by Hansen & Sato and Annan & Hargreaves. The ~3C estimate for Holocene conditions is looking increasingly solid.

I've downloaded Thompson & Sieber but I won't read it tonight. I'll comment further tomorrow.

Blast - keep spotting things :

You're pointing me at Zachos et al. (2001). Hansen & Sato (2011) uses Zachos' Cenozoic temperature reconstruction. I thought you knew this? Just read H&S from section 2: Cenozoic climate change. The proposed correlation between CO2 and Cenozoic climate is detailed there.

http://arxiv.org/ftp/arxiv/papers/1105/1105.0968.pdf

"Is that your point - that climate on a world with ice caps is less stable than on an ice-free world?"

Yes, that was my suggestion. How does CO2 explain this behaviour?

My main point however, is that climate sensitivity is not constant. See the papers I've pointed you at over the past few days, as well as papers by Tsonis on modes and synchronization, and papers by Rial on ice age transitions. It seems that CS is constant neither across glacial/inter-glacial transitions nor even during the Holocene. In particular, the argument in H&S11, based on glacial-interglacial changes, is undermined by the non-constancy of CS.

"Hansen & Sato (2011) uses Zachos' Cenozoic temperature reconstruction. I thought you knew this?"

Yes, I do know that, remember I pointed you at the diagram in the H&S paper. Nonetheless, the Zachos et al paper does, as I stated, present the view that there were multiple causes for Cenozoic cooling, including the "potential for highly nonlinear responses in climate to forcing" (in addition to the influence of CO2). IMO, this viewpoint differs from that presented in H&S.

Philip

Yes, that was my suggestion. How does CO2 explain this behaviour?

It's widely accepted that ice-albedo feedback increases instability - with the help of GHGs under Milankovitch forcing. This does not mean that CO2 is not the dominant long-term factor in Cenozoic cooling. H&S11 quantifies the CO2 forcing over the Cenozoic as follows:

In contrast, atmospheric CO2 during the Cenozoic changed from about 1000 ppm in the early Cenozoic (Beerling and Royer, 2011) to as small as 170 ppm during recent ice ages (Luthi et al., 2008). The resulting climate forcing, which can be computed accurately for this CO2 range using formulae in Table 1 of Hansen et al. (2000), exceeds 10 W/m2. CO2 was clearly the dominant climate forcing in the Cenozoic.

Just briefly: Hansen (as previously) has used data for GAT change during the LGM vs the modern period. H&S11 uses conservative estimates for the major forcings that drive the glacial termination (principally albedo and GHGs)

H&S uses the ratio of change in GAT to the total forcing and defines this as the sensitivity. This approach is essentially empirical, as H&S state:

Models are imperfect and we will never be sure that they include all important processes. Fortunately, Earth's history provides a remarkably rich record of how our planet responded to climate forcings in the past. Paleoclimate records yield, by far, our most accurate assessment of climate sensitivity and climate feedbacks.

Yes, it assumes a constant CS between different states, but what's interesting is that you get the same best estimate for CS when you change the approach. Consider eg Schneider von Deimling et al. (2006) where ensemble simulations are used to constrain paleo-data. This perturbed physics approach does not require a constant CS between the LGM and the Holocene. You still end up with a CS estimate of ~3C.

One question did come up from what you say:

It seems that CS is constant neither across glacial/inter-glacial transitions nor even during the Holocene.

Who claims that CS is not constant during the Holocene and why?

I'll respond on T&S as soon as I've had a chance to read it.

Thanks! I think we have some agreement then that sensitivity varies across the span of a glacial/interglacial, and hopefully the understanding that this happens because there are turning points along the way. By the same logic, we should presumably expect variation of S in the Holocene as well, if similar structures have been traversed -- and interaction between different modes has been identified as one possible source (by Tsonis).

Regarding the idea that the ice-albedo effect increases instability, this should presumably also imply that changes caused by a given forcing are smaller if temperatures are higher than currently, as opposed to lower.

Regarding Schneider von Deimling et al, I can only view the abstract, so can't really get much out of it. However, I admit to feeling a little perplexed by the suggestion that their approach doesn't assume a constant sensitivity, and yet still ends up calculating one!

I should also mention that the remarks I've made about sensitivity don't mean I'm taking a position regarding the possible damage due to CO2 emissions -- right now I'm simply interested in trying to understand a little better the implications of the non-linearity. Any other papers you've noticed that might help, please point them out. For example, one currently unanswered question I have is to do with explaining the scaling observed in the Pleistocene temperature record, although I imagine it is related to instability in the vicinity of turning points.

Philip

Going over Zachos et al. again prompts me to place the quote you provide - twice - from it into context. You say:

Nonetheless, the Zachos et al paper does, as I stated, present the view that there were multiple causes for Cenozoic cooling, including the "potential for highly nonlinear responses in climate to forcing" (in addition to the influence of CO2). IMO, this viewpoint differs from that presented in H&S.

From Zachos et al.:

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.

I think the context adds some important clarity here.

SvD et al pdf here, Sorry - my mistake.

http://pik-potsdam.de/~stefan/Publications/Journals/Schneider_etal_ClimDyn_2006.pdf

Philip

By the same logic, we should presumably expect variation of S in the Holocene as well, if similar structures have been traversed -- and interaction between different modes has been identified as one possible source (by Tsonis).

Whoah! Too far, too fast ;-) I'm not aware of any evidence that CS (S) has varied during the Holocene.

You say:

I admit to feeling a little perplexed by the suggestion that their approach doesn't assume a constant sensitivity, and yet still ends up calculating one!

I think there is a simple misunderstanding here. I meant that SvD's approach doesn't require the prior assumption that S is constant in the LGM and the Holocene. Rather that the value for S is an emergent value from model runs constrained with paleo data.

Philip

Thanks for the Thompson & Sieber paper which is interesting although I do not claim to be able to follow the mathematical section closely.

The question here is what relevance does a paper purportedly aimed at predicting non-linear responses have to estimating climate sensitivity? Is it as useful in this respect as Annan & Hargreaves (2006), Hansen & Sato (2011) or Schneider von Deimling et al, (2006)?

For me, the answer is no. Nor, despite considerable effort, can I reconcile with any of your earlier statement that:

I'm doubtful about the meaningfulness of the idea of a CO2 doubling sensitivity, but happy with the kind of transient sensitivity treated by Spencer and Lindzen. With this caveat in place, then inside the linear regime, I'd expect sensitivity to be low as per those authors

You may be doubtful about the meaningfulness of the idea of a CO2 doubling sensitivity, but that does not mean that others are. Nor does reference to Spencer and Lindzen inspire confidence: there is simply no support for the extremely low CS estimates they propose. There is however a large and growing body of work that contradicts such low estimates and points to a most likely value of ~3C (per AR4 WG1).

My sense is that you seek to discount the scientific consensus on climate sensitivity to CO2 because it is at odds in some way with strongly held personal beliefs. I have noticed this in your exchanges on this thread with Richard, as well as variously at BH with me.

I will leave the last word to Tripati et al. (2009) Coupling of CO2 and Ice Sheet Stability Over Major Climate Transitions of the Last 20 Million Years:

Variations in pCO2 affect the radiative budget and energy balance of the planet. Such changes will inevitably have consequences for temperature, the hydrologic cycle, heat transport, and the accumulation and ablation of sea ice and glacial ice. The data presented here do not preclude alternative mechanisms for driving climate change over the past 20 Ma. However, they do indicate that changes in pCO2 were closely tied to the evolution of climate during the Middle and Late Miocene and the Late Pliocene glacial intensification, and therefore, it is logical to deduce that pCO2 played an important role in driving these transitions. High-resolution records of pCO2 and other climate parameters should help to resolve whether pCO2 was a trigger and/or feedback (or both).

These results provide some constraints on pCO2 thresholds for the advance and retreat of continental ice sheets in the past, which is also relevant in the context of anthropogenic climate change because it is uncertain how continental ice sheets will respond over the coming centuries to increased levels of pCO2 (1). By comparing our reconstruction to the published data sets described above, we are able to estimate past thresholds for the buildup of ice in different regions. When pCO2 levels were last similar to modern values (that is, greater than 350 to 400 ppmv), there was little glacial ice on land or sea ice in the Arctic, and a marine-based ice mass on Antarctica was not viable. A sea ice cap on the Arctic Ocean and a large permanent ice sheet were maintained on East Antarctica when pCO2 values fell below this threshold. Lower levels were necessary for the growth of large ice masses on West Antarctica (~250 to 300 ppmv) and Greenland (~220 to 260 ppmv). These values are lower than those indicated by a recent modeling study, which suggested that the threshold on East Antarctica may have been three times greater than in the Northern Hemisphere (35).

This work may support a relatively high climate sensitivity to pCO2. pCO2 values associated with major climate transitions of the past 20 Ma are similar to modern levels. During theMid-Miocene, when pCO2 was apparently grossly similar to modern levels, global surface temperatures were, on average, 3 to 6°C warmer than in the present (2, 25).We suggest that the Mid-Miocene may be a useful interval to study to understand what effect sustained high pCO2 levels (i.e., a climate in equilibrium with near-modern pCO2 values) may have on climate.

"My sense is that you seek to discount the scientific consensus on climate sensitivity to CO2 because it is at odds in some way with strongly held personal beliefs."

If you're looking for something nasty, I'm afraid you won't find it. I definitely do want to get a better understanding of the likely effect of increased CO2, and because of the politicization of the science, this means I think it is important to try as hard as possible to strip down the arguments, rather than simply accepting what I'm told. In particular, it means understanding the impact of the non-linear dynamical effects. It's obvious (at least to me) that there is a very poor understanding of this aspect, even amongst professional climate scientists. If you think that this observation is wrong, and you want to change my mind about it, then you will need to demonstrate that it is wrong, by showing me evidence of a good understanding. I think you'll struggle to do this!

Perhaps it might help you to understand my position a little better if you go through the exercise of writing down your understanding of the definition of climate sensitivity. I notice that you either missed or accepted the definition I suggested earlier (dT/dF - the slope of the curve of T vs F, as exemplified by Z&G figure 5). If you think about it for a moment, you may want to argue against this suggestion, and replace it with something else ... but what?

Regarding variability of CS during the Holocene, I again point you to the synchronization argument of Tsonis; to the various examples of turning points in T&S; and also to the examples in Rial's review paper. I'd certainly agree with you if you state that none of this is conclusive. However, the text of Z&G makes it plain as a pikestaff that it is a reasonable starting point to assume that CS varies in the way I've suggested (unless "CS" means something rather different to dT/dF). Again, if you pause, the variability of CS (if true) provides a simple explanation for the problems in pinning its value down more closely.

To respond to some of your other points:

Tripati et al: There's nothing in your snip I'd particularly object to. I certainly agree fully with the first two sentences. Thanks for the reference!

SvD et al: From the abstract they conclude that CS lies in a range from 1.2 - 5.3 K. This seems fair enough to me. I will try to read it carefully later on, but since my training is in physics, I am only able to read one word at a time (if that)!

Zachos et al: I take it that your point is that you think they mean nonlinear responses to forcing only occur when there are significant ice caps? If so, I'd question this interpretation, although I agree their wording is somewhat ambiguous. The reason I take their comment in the way that I do probably reflects my own context on the issue i.e. nonlinear climate effects must have occurred throughout earth's history, indeed their statement starts with the words, "As for the aberrations, there mere existence points toward ... ". You might also like to reconsider The Pliocene Paradox paper -- start of the abstract, "During the early Pliocene, 5 to 3 million years ago (Ma), globally averaged temperatures were significantly higher than they are today even though the external factors that determine climate were essentially the same."

Spencer and Lindzen: "there is simply no support for the extremely low CS estimates they propose." At the risk of stating the obvious, their papers present the evidence for extremely low CS (although I accept that you need to interpret this as meaning something along the lines of the instantaneous value of dT/dS - but I know that Spencer at least also accepts something along these lines because I asked him on his blog; I don't know Lindzen's position on interpretation).

Richard Betts: I can assure you I have the highest regard for Richard, both for his expertise and for his willingness to engage with people here. I give him a hard time sometimes because I think a lot of the arguments he's presented to us are faulty or flat out wrong. I think this results from a tendency to pass on arguments outside his specialty that he's heard from others, without really bothering to carefully check for himself how reasonable they are (but that's not particularly a fault - it's the way science normally works!). All of this doesn't necessarily mean I disagree with his conclusions. Do you want me to quietly accept his arguments when I am convinced they are incorrect? I certainly hope that Richard wouldn't want me to do that, and I trust that he will take it in the same spirit if I carry on being critical!

Zachos et al: I take it that your point is that you think they mean nonlinear responses to forcing only occur when there are significant ice caps?

Seriously, is this even part of the thinking? To what extent does anyone who spends their life in their lab, rather than out on a boat, think such a difference in albedo is possible?

Is there some ineluctable relationship to a world without wind?

I can tell you from living beside the Waitemata - "sparkling waters" - about the reflective capability of the gentlest wind-lapped water. The glare is blinding. It's really not like the grey North Sea or the Hamptons or Northern California. And has no-one considered the term "White Cap"? 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.

Gixxer,

Possibly you have misunderstood - the above is not a quote from the Zachos paper, but a response to BBD's earlier comment @Jan 17, 2012 at 3:48 PM. The argument is to do with the life-cycle of the ice-caps, when they melt and when they form. Apparently, melting in particular can occur rather quickly, presumably the result of a non-linear response. I'd suggested earlier that such responses can occur even when there isn't significant glaciation (e.g. now), and that the Zachos paper supports this idea. BBD, I think, disagrees on both points.

Thanks Philip

Indeed, I might well have misunderstood. Apologies if I have.

It's just I have been reading much of late about non-linear CS, and albedo keeps coming up as an issue. Along with planetary black bodies, grey bodies, Stefan-Bolzmann and 'emissivity'.

Fine and dandy, but the assumptions that go into these (mathematically impenetrable for us Arts Grads) equations look redick. Like the whole radiative '1' for a black body but, hey, "More likely 0.95 for a real rock." And the assumptions about albedo relating to ice vs water - where did that come from? I've read estimates of 0.5. Really? And as I pointed out, wind can turn the water surface 50% white. And old ice is often dull grey - less reflective than water.

This might all be factored into the models but all I read about are hypothetical assumptions and averages. A simple smell test against real world experience looks completely wrong.

?