Okay, not really, but I might need to change them. Something that I’ve been stressing here is that internal (natural) variability can’t produce long-term warming. The reason being that it just moves energy around the climate system. If it’s not associated with some kind of change in radiative forcing, then if it heats the surface, the extra energy will be lost quickly, and if it cools the surface, the energy will be recovered quickly. For a system already in equilibrium, internal variability will simply cause the temperatures to vary about the externally determined equilibrium temperature.
I have, however, been made aware of a recent paper (thanks Victor; some people just don’t know how to stick to the script) that shows that it may be more complex than this. The paper (Top-of-atmosphere radiative contribution to unforced decadal global temperature variability in climate models by Brown, Li, Li & Ming) uses 36 pre-industrial control runs from the CMIP5 ensemble, to investigate how decadal variability in the top-of-the-atmosphere (TOA) flux influences surface temperatures. As I understand it, all these runs are unforced, so all the variability is internal (natural) and not forced.
The main figure is probably the one below. The two largest-magnitude warming decades and the two largest-magnitude cooling decades were selected from each of the 36 controls runs. In each case, they also determined the TOA flux for each of the decades that were selected. What they found was that warming decades were associated with a mean net TOA flux of -0.06 Wm-2 (upward orientated, so negative means gaining energy), and the cooling decades were associated with a mean net TOA flux of +0.06 Wm-2. So, even though this variation is internal, the warming is not simply associated with a movement of energy around the climate system, but is also associated with a net TOA flux that increases the total energy in the system. Similarly – but in reverse – for cooling decades. The paper also illustrates that about half the warming (cooling) in each of these decades was associated with the net TOA flux.
Figure showing temperature trends and TOA fluxes from unfroced control runs (Brown et al. 2014).
So, this seems to be an interesting result : internal variability isn’t necessarily simply associated with energy moving around the climate system, it could also associated with a net TOA flux. The paper indicates that this is mainly a consequence of small changes (0.1%) in the albedo that temporarily counteracts the change in the outgoing long-wavelength radiation.
The one question, though, is how significant this is with respect to recent warming. The paper addresses this, saying,
From the years 1955 to 2010, 0–2000 m ocean heat content accumulation suggested that the mean QTOA over this period was ~0.27 W/m2 (due mostly to external radiative forcings, F) [Levitus et al., 2012]. This study has shown that in extreme episodes of decadal scale unforced T change, mean TOA imbalances were on the order of ± ~0.06 W/m2 averaged across all GCMs (the most extreme imbalances observed were ± ~0.2 W/m2, Figure S4). This would imply that in certain circumstances, unforced variability in QTOA may have been able to modulate the long-term forced imbalance by ~22% (~74% in the most extreme circumstances) over the course of a given decade. Currently, however, measurements indicate that the energy imbalance at the TOA is between 0.5 and 1.0 W/m2 [Abraham et al., 2013; Trenberth et al., 2014]. At this magnitude, unforced QTOA variability would only be able to modulate the background forced imbalance by ~6–12% (~20–40% in the most extreme circumstances) over the entirety of a given decade.
So, over the course of a given decade it could modulate the forced imbalance by more than 50% when the imbalance is less than 0.5 Wm-2. However, given that the imbalance today is probably around 0.5 Wm-2 or greater, even in the most extreme scenario, internal variability will probably only be able to modulate the imbalance, today, by about 40%, at most. Of course, one should remember that this could both increase and decrease the imbalance, and – if I understand the paper correctly – is unlikely to extend beyond about a decade.
Even though this paper indicates that internal variability may not be only associated with moving energy around the climate system, I don’t think this means that internal variability can – by itself – be associated with long-term (multi-decade) warming, or cooling. I thought I might lay out some reasons why. Of course, these are just how I understand it, so if others disagree or want to elaborate or clarify, please do so.
- Unforced, climate model control runs do not show evidence for multi-decade trends associated with internal variability.
- Even though internal variability may not be only associated with moving energy around the climate system, this is still the process that triggers the warming or cooling. Since the available energy is finite, the process is self-limiting.
- There is little evidence to suggest that internal variability played a significant role in past climate variability. Maybe, more correctly, in most cases we can associate past changes to our climate with external triggers or changes in external forcings.
- As I understand it, the main reason for the change in the TOA flux is because of changes in albedo, presumably driven by changes in cloud cover. One might think that this could feedback on itself and consequently produce longer-term warming or cooling. Water vapour is, however, quite sensitive to temperature changes and so when the cycle reverses, the changes in water vapour and clouds should reverse and the system should evolve back towards to its pre-warming/cooling state (I may not have explained this all that well, so if someone can do a better job, go ahead).
- The annual variation in global temperature is actually greater than the kind of variations being considered here. If this kind of variability could produce long-term warming, then that would suggest that the same should apply to the annual variations.
- If internal variability can drive long-term warming, it would suggest that our climate is quite sensitive to small changes. This should then apply to changes in external forcings too (increased CO2 concentrations, for example), but might also imply that if it were that sensitive to small changes, we wouldn’t be here to have this discussion.
Given the recent interest in the role of internal variability, I thought this might be something interesting to post about. I don’t, however, think it really changes anything with respect to the significance of internal variability (and it is only one paper) and have laid out some ideas why, above. Of course, if others disagree or have other thoughts, feel free to make them through the comments.
...and Then There's Physics 
I don’t think things are as simple as the paper likes to present it, the “random TOA flux” value for one decade is very likely to be VERY different than the value over any 4 consecutive decades. On top of that under normal conditions there is a certain TOA forcing from the solar cycle, and that one is likely in the order of 0.5W/m^2, so even on a single decade it could dwarfs the “random” 0.06 found by the models, and on 4 decades I would be surprised if the trully random TOA flux would remain 0.01. But that is just a guess.
Catalin,
Certainly, and as I understand it, it is very unlikely that these changes could be sustained for multiple decades. Also, as you say, if it is typically 0.06Wm-2 or smaller, it will probably be dwarfed by other factors that will also be causing the TOA flux to vary (hadn’t actually thought of that). I think it is possible for it to be larger than that, but that then becomes much less likely.
Anders… some time back I think I saw a quote on Skeptical Science which mentioned a foot note from a paper saying that the simulations were averaged. And many of the simulations showed decade long pauses and even declines occurring.
I haven’t been able to find it… Can you recall?
AoM,
Not off the top of my head, but I thought this was the basic issue. If you’re referring to the forced runs that try to model the 20th century and early 21st century, what’s normally presented are ensemble averages. Since different models can’t exactly predict the same variability, that gets averaged out. So, it’s my understanding that models do indeed predict decade long pauses, but just can’t predict precisely when they’ll occur.
That’s it. The implication is of course that this is already occurring in the models, and therefore they are right.
That’s actually not a terribly surprising finding although as usual numbers are better than intution.
Dithering around zero or so goes nowhere, unlike our own reliable manufactured warming.
Doug,
Yes, I was quite interested to see this. As you say, maybe not that surprising, but it is better to actually show it, rather than just guess.
Andthentheresphysics
You say
“Something that I’ve been stressing here is that internal (natural) variability can’t produce long-term warming. The reason being that it just moves energy around the climate system..”
I’ve seen statements of this class a lot, including sometimes from very knowledgeable climate scientists and it is an unsupportable claim.
I’d like to demonstrate the flaw in this thinking clearly on Science of Doom at some stage. Who knows if I will succeed in my goal.
But for now.. a non-linear system can exhibit variability that is “unexpected” from the drivers that are forcing it. The simple Lorenz 1963 example – shown in Natural Variability and Chaos – Two – Lorenz 1963 is probably the best known.
Well-known physics professor Pierrehumbert gives a great example in his 1995 paper. I highlighted this paper in Clouds and Water Vapor – Part Five – Back of the envelope calcs from Pierrehumbert.
Linear systems will definitely do what you say. Some non-linear systems will do what you say. Many will not.
In the deity-like position of being able to see the complete attractor space of the Lorenz system we can say that “we know the statistics of this system”, a possible analogy being climate vs weather. In the non-deity like position of viewing some part of the time-series we might be right on the money with that last statement, or we might be hopelessly flawed.
If you were correct that causes of past climate change are well-understood your statement might be proven accurate. On this working assumption, climate models might then be used to determine the accuracy of your statement. Even under this assumption you might well be wrong because non-linear interaction of sub-grid processes can produce unpredictable changes.
I know that many climate scientists say “we know what caused the ice ages”. And I see that the IPCC AR5 also holds this opinion. In my unscientific survey of the 300 papers in my ice ages folder I found that half of those who ventured an opinion on the current state of the field (not their own, about to be revealed, new idea) thought that the causes of ice age terminations were “unknown”. Many said “it is widely believed” which is a fascinating insight into our understanding of ice age terminations. Climate modeling of terminations has not had success and climate modeling of inceptions is at a preliminary state.
There are a lot of other faster climate transitions, where the causes are still not understood. This doesn’t mean people haven’t put forward ideas of course.
SoD,
Sure. However, I do think that I have mostly qualified that statement when I’ve used it (apart from here, which was intentional as the rest of the post was about how it’s wrong). This post was intended to be one showing that it isn’t correct – as I assume you realised.
I think you may be putting words in my mouth, as I thought I’d been quite careful as to how I described past climate changes. I don’t believe I said that they’re well understood. This is actually what I said
which – I would argue – is a much weaker statement that what seem to be suggesting I said.
Just to be clear, I agree with your comment. This post was – to a certain extent – an acknowledgement that some of what I’ve said previously may have been too strong (although I do think I typically qualify these statements). I do agree with you about non-linear systems, but my understanding is that – at the moment at least – there is little evidence to suggest that our climate is sensitive to small internal perturbations. That doesn’t mean that we should assume that it isn’t, but – similarly – doesn’t mean that we should assume that it is. This paper seems like an interesting step at trying to quantify it’s significance. It’s not the last word, obviously, but a step.
I’m aware of the D-O events that are potentially unforced so, yes, I agree. Again, though, there’s a difference between us not knowing if something is possible and assuming that it is.
I’ll make a comment about this
In a sense scientists should be objective and not hold views that aren’t entirely supported by the evidence. They shouldn’t say something that they can’t back up completely. In reality, however, scientists often do have views that they can’t back up completely. Sometimes this is because they have a bias in favour of some pet theory and they’ll turn out to be wrong. Other times, however, it’s because they (and others) have spent decades working on something and understand the system well enough to hold a view that maybe they can’t fully support, but that is unlikely to be wrong. I agree that we probably can’t say that we “know” the causes of the terminations, but if a large group of experts feel that they essentially do, one should possibly accept that they might have a point. They may turn out to be wrong of course, but it is useful to know what there views are and one shouldn’t completely dismiss them if they do hold such a view.
> many climate scientists say “we know what caused the ice ages”. And I see that the IPCC AR5 also holds this opinion
Having made mistakes on your threads I’m reluctant to contradict you, but that sounds implausible. Section 5.22 (http://www.climatechange2013.org/images/report/WG1AR5_Chapter05_FINAL.pdf) says “Orbital forcing is considered the pacemaker of transitions between glacials and interglacials (high confidence), although there is still no consensus on exactly how the different physical processes influenced by insolation changes interact to influence ice sheet volume (Box 5.2; Section 5.3.2)” and your paraphrase doesn’t rally agree with that; perhaps; you were thinking of a different section.
“If internal variability can drive long-term warming, it would suggest that our climate is quite sensitive to small changes. This should then apply to changes in external forcings too (increased CO2 concentrations, for example), but might also imply that if it were that sensitive to small changes, we wouldn’t be here to have this discussion.”
If it’s as simple as climate feedback is a constant then yes.
Obviously there are positive feedbacks in the climate. For example, the indisputable ice albedo feedback. Obviously there are negative feedbacks. How do they interact and change?
Is the “feedback” constant through space and time? I suggest extremely unlikely. I can’t prove it. Of course, I think the climate record of the last few hundred thousand years backs up this position.
Graeme Stephens (2005) has a good comment on the “constant linear feedback” concept:
“..While we cannot necessarily dismiss the value of [the eqn of constant linear feedback] and related interpretation out of hand, the global response, as will become apparent in section 9, is the accumulated result of complex regional responses that appear to be controlled by more local-scale processes that vary in space and time.
If we are to assume gross time–space averages to represent the effects of these processes, then the assumptions inherent to [the eqn] certainly require a much more careful level of justification than has been given.
At this time it is unclear as to the specific value of a global-mean sensitivity as a measure of feedback other than providing a compact and convenient measure of model-to-model differences to a fixed climate forcing..”
– Cloud Feedbacks in the Climate System: A Critical Review (2005)
William,
I read “..Orbital forcing is considered the pacemaker of transitions between glacials and interglacials (high confidence)..” as “we know what causes ice ages to start and end” and the caveat “although there is still no consensus on exactly how the different physical processes” as “just some details that remain to be sorted out”.
If IPCC AR5 is saying that ice age terminations are still a mystery then I’m delighted to be so well-aligned with this august group.
The phrase ‘internal variability’ also tends to gets thrown around a little to casually without being well defined.
Internal variability that is just internal to the atmosphere needs to be thought of quite differently from ‘internal’ that includes all the other parts of the system – oceans, cryosphere, land and biosphere.
The atmosphere has so little thermal mass that even small energy imbalances can quite quickly change atmospheric temps and thus quickly vary radiative forcing. So conversely any internal variability that is mainly intra-atmospheric in origin will of necessity be short lived because there isn’t the thermal mass needed to sustain it.
Once the oceans particularly become involved, contributing their thermal mass it is a whole different ball game. ENSO is a short term example of this. In an El Nino year net heat flows from the oceans to the atmosphere, and there it increases outgoing LW flux as temperatures rise. And this can be sustained for as long as the El Nino conditions persist because the oceans have the thermal mass to sustain it. And La Nina has the opposite effect.
So perturbations in flow patterns in the oceans could quite certainly last for longer periods – decadal scales or so. And as a consequence fluctuations in OLR on those timescales as well.
That is why the data on Ocean Heat Content is such central evidence for the warming. If all we had was atmospheric temperature data we wouldn’t be able to rule increased heat flows from the oceans as a cause.
Climate really is a dog called Ocean. We just live on the tip its tail.
ATTP- I don’t have access to the full paper, but you say “The paper also illustrates that about half the warming (cooling) in each of these decades was associated with the net TOA flux.” Since these are unforced runs, where does the other half come from? The reason I ask is that if an unforced temperature change – e.g., warming – is half associated with increased TOA net heat flux and half associated with heat release from the ocean, then the net result is expected to be a net heat loss from the ocean. If the ocean is actually gaining OHC during a warming decade, the internal contribution, while not necessarily close to zero, should be very small compared with the forced component, which since 1950 has been almost exclusively anthropogenic. The dominance of forced warming post-1950 or from 1975-2000 appears to be affirmed.
In addition, if one is looking at a particular historical period such as 1975-2000, any significant internal contribution to the warming based on model runs must match the observed geographical distribution of the warming. Isaac Held has discussed this in detail at heat uptake and internal variability.
Addendum – For those who like me, don’t have access to the full paper, the following presentation of its main findings may be useful – Brown 2014 This article appears to confirm the inference that unforced variability (at least in the model runs) is unlikely to mediate both warming and a concurrent increase in OHC. The latter, if observed, implies a dominant forced component as quantified in Isaac Held’s discussion – see the link in my above comment.
Another try at the link to Brown – Brown 2014
SoD,
I agree. I think my point was more simply that if our climate is extremely sensitive, the how has it appeared to remain so stable over such a long period of time. Of course, that doesn’t mean that it will always be so.
Fred,
Yes, that’s how understand it. The process is driven by the movement of energy, but amplified by a net TOA flux. So, yes, it would seem that this variability should typically be associated with a change in OHC.
Fred,
Thanks for the link. I hadn’t seen that yet.
Has our whole interaction on this topic disappeared down the memory hole, SoD? Among other things, Abe-Ouchi et al. (2013) was too late for the AR5. You need to align yourself with the present.
ATTP,
It’s essential to remember that this paper is totally about present models, not directly about the real Earth system. I for one do think that this kind of variability is not necessarily well described by the models.
If a phenomenon takes place in a model and isn’t due to feature known to be wrong, it can be concluded that it’s possible, i.e. it doesn’t contradict, what we know. The contrary is, however, not true: That something is not caught up by models, it may still be true in the real Earth system. I do not believe that general principles set stringent limits at all for the strength of variability in either the heat uptake of the oceans or in the TOA balance. Observational data has its own limits.
In my view you continue to make too strong statements on the issues. My proposal is:
There are great uncertainties about the internal variability, but emphasizing the effect of that uncertainty on the downside of climate sensitivity represents wrong logic in the message to decision makers.
Pekka,
Indeed, that’s a fair point.
Indeed, I do realise that you think that. I get the impression that maybe you’d like me to become more like Judith, which is highly unlikely to happen. Although this is consistent with what you say, it seems that what you interpret me as saying is too strong, rather than what I’m saying actually being too strong.
Okay, I probably agree with that, but I don’t think that I’ve mentioned policy makers much, so you’re extending things beyond what I’m trying to say.
Steve,
While the work of Abe-Ouchi et al agrees very well with, how I like to see the situation, I do not believe that its ideas have been confirmed nearly well enough. It’s the latest proposal (or one of the latest) and still far too young for being considered well established scientific knowledge.
The problem with all paleoclimatic work is that the approach is highly dependent on narratives. The pieces of information are too sparse to define the full picture of the past state of the Earth system and in many cases even for telling precisely what a particular observation really means. Adding a narrative to that makes a full picture. If the full picture avoids serious internal contradictions, it may be correct, but it need not be. As the number of competent people who create independently such narratives is very small, the risk that something essential is missed or in error is substantial.
When there are, in addition, some contradictions between the narrative and some observations, the risk of errors just grows. Often the error is in the conflicting piece of data, but it may well be also in the narrative.
I think that we can still say that internal (natural) variability can’t produce long-term warming if we define “warming” to be the accelerating warming in the oceans, if we always stick to including the oceans.
It seems to me that the deniers want the masses to ignore the oceans – it makes their denials look better when they don’t have to explain where all that accumulating heat is coming from. I mean, they keep trying to say global warming stopped in 1998 even though since then, heat has been accumulating in the oceans at an accelerating rate for an average over that time at the rate of roughly 1/3 million Hiroshima-sized nuclear bombs per day.
I’d like to call everyone’s attention to the most recent post at “Climate Consensus – the 97%” by Dana Nuccitelli, where he at least here gets it right by including the term “surface”, that global warming has not stopped, has not stopped its acceleration upward:
“Unpacking unpaused global warming – climate models got it right: Global surface warming has slowed down due to internal and external factors, consistent with climate model predictions that account for these effects”
http://www.theguardian.com/environment/climate-consensus-97-per-cent/2014/aug/25/unpacking-unpaused-global-warming-climate-models-right
Is it not true that there are generally only three ways to heat the oceans (where more than 90% of the energy does its thing) and atmosphere (where only roughly 2% of the total energy does its thing) long-term, these being (1) more heat falling on Earth from the sun, (2) more of that heat falling on Earth being absorbed (lower albedo), and (3) more of that absorbed heat being trapped by the greenhouse gas effect? And is it not true that the first two cannot explain this long-term accelerating heat accumulation and that therefore by the process of elimination the answer must be behind door number (3)? With respect to why the entire system is warming at an accelerating rate, it’s all so simple, and this simplicity I think should be emphasized to the public without end.
Trying to understand how much of what’s happening at the surface is natural is important, but for the sake of winning the hearts and minds of the public, it seems to me that it should always be done in such a way that never loses the focus on where almost all the energy is and that its accumulating at an accelerating rate.
Just for the sake of argument, as to what to say to the masses that might be confused by denier arguments: Suppose every last bit of the recent surface warming was “natural” in that it came from the oceans. Would that therefore mean that spewing all these greenhouse gases into the environment won’t hurt? No, it would still hurt – in the long run, too much heat will destroy future civilization and much of life on Earth regardless of how much of it first passes “through” the oceans.
Fred,
An interesting aspect of the post by Patrick Brown on Ed Hawkins’s blog is what he says about the hiatus. Their models would suggest that if there is a cooling trend relative to the forced trend, you would expect to see a reduction in the net TOA flux (i.e., we should be accruing less energy than in periods where there is no cooling trend). The point they make about the hiatus period is that although the trend appears slower than we were expecting, there does not appear to be a reduction in the net TOA flux. I guess it could be that it might have been even higher had we not been in a cooling phase, but there doesn’t appear to be a drop in the net TOA flux.
On a similar note, I saw what appeared to be a remarkable tweet from Roger Pielke Sr yesterday
He’s referring to Dana saying that The global climate has continued to build up heat at an incredibly rapid rate and yet has chosen to highlight the 0-700m OHC. As I understand it, the 0-2000m data shows a continued rise in OHC. I guess whether or not it is incredibly rapid might be subjective (although it may well be faster than at any time in human history) but cherry picking one portion of a dataset and saying that what someone else is saying is “absurd” seems a remarkable thing for a high-profile, senior scientist to say. I did point out the 0-2000m data to Roger Pielke Sr, but he didn’t respond.
KeefeAndAmanda,
That Guardian article of Dana’s is what I think Roger Pielke Sr is referring to in his tweet.
I think when a senior scientist starts tacitly asserting the oceans are only 700 meters deep, we can safely so he has gone emeritus.
Looking at the Fig. 1 of the Chen et al paper you posted about a few days ago we can see that the increase in the layer between 700m and 1500m over the latest years seems to come mostly from the Southern Ocean and over the previous years (2000-05) from the Atlantic.
I’m a little uncertain on the reliability of the overall numbers, when the changes are so localized. Rather small coverage issues might distort the results, when the changes vary so much from one part of the oceans to another, and there are still issues even with the present ARGO, and more before (including the first years of ARGO).
Pekka,
Sure, but there’s a huge difference between been uncertain about the numbers and claiming that what someone has said is absurd (especially if you base that on a subset of the data). My understanding is that there is very little chance that the OHC has not risen in the last decade or so and also that it seems quite likely that it is currently rising faster than its been in previous decades.
ATTP – “My understanding is that there is very little chance that the OHC has risen in the last decade”
Typo alert!!!
Rob,
Indeed, thanks. Forgot the word “not” between has and risen. Fixed now.
Steve Bloom,
“Has our whole interaction on this topic disappeared down the memory hole, SoD? Among other things, Abe-Ouchi et al. (2013) was too late for the AR5. You need to align yourself with the present.”
No, but our interaction was not a fait accompli on the topic. Pick a year, pick a result from a climate paper where voila, the problem is solved. Pick another paper the same year, or a different year and there is a different solution.
I have read Insolation-driven 100,000-year glacial cycles and hysteresis of ice-sheet volume, Abe-Ouchi 2013 and also the necessary precursor paper – Climatic Conditions for modelling the Northern Hemisphere ice sheets throughout the ice age cycle (necessary because the 2013 paper references this for all details on the model). I wrote one article about this earlier paper – Ghosts of Climates Past – Nineteen – Ice Sheet Models I.
At the risk of over-simplifying my summary (more in my article), ice models are quite rudimentary and have a lot of parameters with little constraint on them – because the subject is a very non-linear one with a lot of research still in progress.
The complete model in 2007 and 2013 is essentially an ice sheet model with some parameterized inputs from prior run GCMs under certain conditions. The ice sheet model has a lot of unknowns.
So is the result interesting? Yes. Has it developed an older hypothesis? Yes. Has it developed a strong case for the cause of termination I? I don’t think so.
Here are two main problems to be resolved:
1. Why did the southern hemisphere warm up in advance of TI (the last ice age termination). Perhaps a fatal blow to the theory and not mentioned by Abe-Ouchi.
2. How can all the free parameters be constrained. For example, here is a figure from the 2007 paper – not sure if I can insert an image, here’s my attempt:
[If it doesn’t appear check out my figure 4 in the article cited]
Different reasonable choices of parameters lead to very different results. Ice sheet parameters have a lot more questions about them.
Let’s put it like this – when you know the result, getting a model to reproduce it is interesting but not a slam dunk. The model can get lots of different results that are equally in accord with the understood physics but don’t get the result that matches history. One solution – the parameters selected were therefore correct. Another solution – the model has too many free parameters and so proves nothing. How to choose between these extremes? How well the same model predicts other things that have other constraints. And remember the history of the southern hemisphere is asking an important question.
ATTP, that was just a small test. sorry about that. 🙂
Now I know you do not only claim to be a “professional and active scientist who teaches and carries out research at a university – regarded by some – as amongst the best in the world”, but that you are actually one. You invested your precious time to try to understand a paper that goes contrary to your opinion, an opinion you have voiced over and over again in public, the paper is actually able to convince you and you write a public post clearly stating you were wrong.
You are a scientist! Chapeau.
SoD, climate change is very small relative to the spatial variability and diurnal and annual cycle. Thus for most processes on small spatio-temporal scales, the response is likely nearly linear, even if those processes themselves are nonlinear.
The processes where a nonlinear response may be possible would thus be the very long term ones, where climate change becomes significant relative to the much smaller variability at these time scales. That would include the ice sheets, vegetation and thermohaline circulation. Hopefully we will solve this problem before these long time scales become important.
I read the quote from the IPCC report as stating, we are sure orbital forcing plays a role (they probably assumed the reader would know that that is much too small and can never be sufficient; they hint at this with the word “pacemaker”/trigger), but that there is no agreement about the rest of the dynamics.
Out of curiosity, Pekka, can you point to a actual problem with the paper other than newness? While I’m happy to discuss any such, my reason for raising it here was SoD’s odd failure to acknowledge that it even exists (“Climate modeling of terminations has not had success (…)”). Re “narratives,” I’m sorry if some of these paleo results constrain your fun.
Also, let’s not forget the context. Despite all of SoD’s “unknowns,” when scientists first started grappling with this problem 40 years ago it was entirely obvious that the timing of the post-MPT glacial cycles was keyed to eccentricity. The details of the mechanism were tricky, but that initial assumption was borne out. So I’m a little mystified as to the point he’s trying to make.
AFAIK his “climate modeling of inceptions is at a preliminary state” is arguably true, but I think to imply that it’s a major problem is wrong. When CO2 gets low enough, it likes to snow. When summer insolation is then reduced sufficiently, the snow fails to melt all the way and begins to accumulate at lower latitudes. Over time, the growing ice sheet provides a cooling feedback that allows it to grow to even lower latitudes and survive periods of increased summer insolation. Not really complicated IMO.
The reason for the MPT is more interesting, but at this point it seems pretty clear that it’s down to a slight CO2 reduction (probably due to the continuing supercontinent cycle), eastern Canadian regolith depletion or a combination of the two.
Summing up, all of that leaves us with, well, not much of importance to solve, really.
Re varying sensitivity under different climate conditions, Hansen and others have done some relevant work. It’s a serious concern for the future.
RP Sr.’s complaints are amusing given that he’s now getting the joules metric he was whining about for years. It’s almost as if he had another agenda…
I don’t see the relevance. The theory is about terminations in the NH. Offhand I can think of a couple of reasons why the SH might have behaved like that, but I’d need to review the details. Reference? And of course we would want to examine the data for all of them, not just the last one.
Re the lapse rate, I would think it would make sense to try to make the model work with the smallest reasonable choice. Too large of a lapse rate would start giving early terminations. Too small and the ice-elevation feedback would lose efficacy. Since we know a lot about what the terminations looked like (insolation conditions, location and volume of ice sheets, rapidity of retreat) are those parameters really so free? If we had a skipped interglacial in the mix it would be a different story, but we don’t.
Let’s not forget CO2, Victor.
Victor,
“SoD, climate change is very small relative to the spatial variability and diurnal and annual cycle. Thus for most processes on small spatio-temporal scales, the response is likely nearly linear, even if those processes themselves are nonlinear.”
Thus?
Can you talk me through the second sentence and how it is obvious from the first sentence. I can’t see the connection. Perhaps defining “small” and “nearly” would be a help.
“The processes where a nonlinear response may be possible would thus be the very long term ones, where climate change becomes significant relative to the much smaller variability at these time scales. ”
I’m struggling again with “thus”. Perhaps it’s really simple but I don’t see the relationship.
And what is “very long term”? 1000 years? 10,000 years? 100,000 years?
Steve Bloom:
“I don’t see the relevance. The theory is about terminations in the NH. Offhand I can think of a couple of reasons why the SH might have behaved like that, but I’d need to review the details. Reference? And of course we would want to examine the data for all of them, not just the last one.”
The relevance is high.
If ice sheet dynamics in the high northern latitudes triggered the end of the last ice age, then southern hemisphere warming should follow northern hemisphere warming. The converse – the reality – indicates that something in the southern hemisphere led the globe out of the ice age.
Reference – the problem has been identified and discussed (without any consensus on a solution) for a long time. If you like I can spend a couple of hours and dig up 10 references to “the problem”.
For now, here is Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation, Shakun, Clark, He, Marcott, Mix, Liu, Otto-Bliesner, Schmittner & Bard, Nature (2012).
I highlighted this paper in Ghosts of Climates Past – Eleven – End of the Last Ice age.
The last termination has the most data around it. Earlier terminations have less data. Even the dating of them is not so clear, with dating conflicts depending on whether you accept the sedimentation record of δ18O in deep ocean cores in the shells of foraminifera or the radiometric dating of corals and caves.
Steve Bloom,
“Re the lapse rate, I would think it would make sense to try to make the model work with the smallest reasonable choice. Too large of a lapse rate would start giving early terminations. Too small and the ice-elevation feedback would lose efficacy. Since we know a lot about what the terminations looked like (insolation conditions, location and volume of ice sheets, rapidity of retreat) are those parameters really so free? If we had a skipped interglacial in the mix it would be a different story, but we don’t.”
It’s a prescribed lapse rate. Unlike the real climate. “Too large of a lapse rate” – with this ice sheet model tied to the parameters input from the separate GCM run – “would start giving early terminations”.
So does that mean we can use the results from this model to work out what the lapse rate actually was? Or question the model. There are a whole set of other parameters. The rebound, the sliding, the “desertification”.. I read the basis papers by SJ Marshall and others. You can pick from a range, and yes, they are “free”. We don’t know what these values are.
This isn’t a GCM producing results. This is an ice sheet model with a lookup table, where the lookup table was produced from a GCM. So it isn’t interactive in any sense. Lapse rates – in this paper – aren’t constrained by anything as good as current convective parameterization in GCMs.
SoD, read the paper again. You’re quite wrong about its subject. Your point about the SH is interesting enough, but not relevant to the paper.
To make a point that should be obvious, the North American ice sheet was key to the end of the glaciations since it contaIned such a large majority of the ice that needed to melt. The paper also addresses the Fennoscandian ice sheet, but it’s much less important due to its much lesser size and the interesting fact that it doesn’t behave in the same manner as the NA ice sheet.
Steve Bloom,
“Despite all of SoD’s “unknowns,” when scientists first started grappling with this problem 40 years ago it was entirely obvious that the timing of the post-MPT glacial cycles was keyed to eccentricity. ”
Interesting.
Entirely obvious to whom?
I’ll pick a few people it isn’t obvious to, with a big publication list themselves:
Obliquity pacing of the late Pleistocene glacial terminations, Peter Huybers & Carl Wunsch (2005)
with the followup
Glacial variability over the last two million years: an extended depth-derived agemodel, continuous obliquity pacing, and the Pleistocene progression, Peter Huybers (2007)
Glacial terminations as southern warmings without northern control, E. W. Wolff, H. Fischer and R. Röthlisberger (2009)
with a nice “simple to digest” conclusion:
“Warming at terminations, as at each Antarctic Isotopic Maximum, is led from the south, but events in the north enable the warming to continue unchecked. The proximal cause of deglaciation is therefore northern conditioning, which paradoxically prevents an early northern DO warming. Although model-based tests are required of our two hypotheses about why the sequence begins (with the cessation of the bipolar seesaw), this work advances understanding of terminations by suggesting that terminations are caused by southern warming that runs away because the north cannot produce a DO event.”
So Wolff and co-authors, Huybers & Wunsch didn’t find it obvious.
I provide more evidence about what climate scientists think in Ghosts of Climates Past – Eighteen – “Probably Nonlinearity” of Unknown Origin.
How would you go about backing up your claims?
SoD, was just a qualitative argument. Helps you find the interesting questions.
Dansgaard-Oeschger cycles are internally forced.
Steve Bloom,
“SoD, read the paper again. You’re quite wrong about its subject. Your point about the SH is interesting enough, but not relevant to the paper.”
You are going to have to help me a little. I’ve read it 5 times. I thought I understood its subject. It doesn’t mention the SH. If that is what you are saying then I agree, the SH is not considered relevant by the authors of the paper.
But it is a relevant question to ask. Evidence not mentioned by a paper that silently asks questions of a paper. Many climate scientists ask this question. All of the “high northern latitude” insolation forcing arguments have been questioned because of (not only because of) the timing of the southern hemisphere warming.
Are the predictions of the paper supported by historical evidence not mentioned by the paper? Or are the predictions of the paper questioned by historical evidence not mentioned by the paper?
Victor,
It was too subtle for me. I like to ask questions because this is how I learn.
Do you want to shed light on your “qualitative argument”?
Wow (tm Judy Curry), SoD, your cup sure is half empty on this one. For one thing, if the focus is just on getting the termination process is it even necessary to have all those other parameters lined up? Re the lapse rate, I don’t understand the issue. The point is that the model works with a reasonable choice. By “works” I mean gets some truly spectacular results. Look at the match to data (i.e. insolation and ice sheet behavior, the first of which isn’t free at all and the second of which isn’t very free). Can you credibly call that just a fortuitous coincidence?
For anyone else who cares about this, the paper is fairly short and readable, and a public copy is here.
I should add that as I am absolutely no expert on this, I verified my assessment by first checking the AGU fall meeting (probably a little too soon to see any pushback, although there was some follow-on work presented), the EGU meeting (plenty of time for pushback but none to be seen, and more follow-on work by a different group), and finally just now (again) by checking to see if any pushback has appeared in the literature given that it’s now been over a year since publication. There’s nothing, and this is an area of research with a fair number of active researchers who in the past have shown no hesitancy to contradict each other. I think if there were major flaws we would have seen something by now.
Richard,
I wasn’t sure that this was definite, but I did mention that in this comment.
Basically I take every scientific paper that presents significantly new results in any field of science with quite a lot of skepticism. I have great trust in the scientific process given enough time, but not in individual papers. SoD seems to have the same basic attitude. In discussion on his site I seem to trust climate models a little more than the other regulars (including SoD), but the difference is not great.
To me the newness of the paper is an essential point. The other question that makes me doubtful on the Abe-Ouchi et al work is the crudeness of the model. On a very general level I have the feeling that the idea must be right, but perhaps not on the details. In addition I try to remember that I must not accept any idea for the reason that it fits so well with my prior thinking. (I mentioned self-criticism in one of my earlier comments. That means specifically that I must remember to be critical on ideas that fit my prior expectations.)
When we are discussing climate and related Earth system components like oceans from the point of view of basic science, the only credible conclusion is that the understanding is in many ways badly lacking and fragmentary. What’s used in weather forecasting is fairly well in control, many specific features important in longer term are also well understood, but many others are not.
People, who are worried about the climate change, and who stick explicitly or implicitly to the linear model (science tells truths, the truths are used in making decisions) do not want to accept that the situation is like that, but that’s not the correct approach. The correct approach accepts fully the uncertainties, and applies approaches of decision-making under uncertainty. In this approach great uncertainties do not prevent reaching strong conclusions, although it makes formulating conclusions rather difficult. That difficulty is, however, due to realities and trying to shortcut it using the linear model simply does not work.
OK, I think I get it. Your claim is that warming from the SH did the job, right? If things are warmer, that’s for sure an extra push, but wouldn’t we want to see retreat not associated with orbital forcing in order to give the SH-driven T increase a major role? Also, what’s plausible for the T increase, and how does its forcing compare to insolation? (I’ll look further at this, but not now since it’s way, way past my bedtime.)
I don’t think any of those folks were ignoring eccentricity, although mention of it may be obscured since it operates by modulating precession. I had a look at Peter’s latest on the topic, and deep into the paper he says “Another implication is that the climate will tend to be driven out
of glaciation sooner when eccentricity—and hence the amplitude of precession—is larger.”
Steve Bloom
I don’t understand this comment: “Wow (tm Judy Curry), SoD, your cup sure is half empty on this one.“.
Is this a “you’re in the wrong camp” argument? I have no camp. If I understand your “argument” correctly, then feel free to ascribe all kinds of sins to my account. And then, like on so many other blogs (of opposite persuasion), not much point continuing a discussion. There is no discussion. I am wrong, because I am not in “the camp”.
As my last summary on your position, I believe my questions about this paper are based on two simple principles. Many other people with a publication record also ask these questions.
1. Did the northern hemisphere ice sheet collapse, as described by Abe-Ouchi, cause the last termination? If so, why was the southern hemisphere warming in advance.
2. With a simple model and many free parameters it is possible to make a model fit the results. A “free parameter” means that the constraint as from physics basics or climate history is not tightly defined. This is the case with ice sheet models where constraints are not well understood.
I call myself a skeptic. This means asking questions and not accepting the argument from authority.
There are arguments from authority and then there’s physics.
Repeating myself: Where did she make such a claim? Or where did I make the claim? Please not to straw man.
This SH stuff is starting to come back to me now. IIRC it got broached about 10 years back and as you say has become well-established.
But it remains that if the NA ice sheet doesn’t collapse the planet is still in a glaciation. A little bit off the top of Antarctica hardly matters.
Note this last sentence of Wollff et al.’s abstract:
Sounds about right.
Pekka,
Sure, that’s a fair, although I think more people do this than maybe you seem willing to acknowledge.
The issue as I see it is that internal variabilities (internal forcings) can drive some climate shifts (D-O events, for example). We also maybe don’t fully understand the mechanism behind the Milankovitch cycles (although it would seem remarkably surprising if it was unrelated to orbital forcings). On the other hand, there seems little evidence to suggest that internal variability can play a major role in the timescales that are of interest to us today, and this paper adds to that basic picture.
Steve & SoD,
I had a quick question. What drives the SH warming?
> I call myself a skeptic. This means asking questions and not accepting the argument from authority.
I call myself a ninja. This means pointing out that “[m]any other people with a publication record also ask these questions” could be taken as an argument from authority.
Offering and accepting an argument may not [be] the same, though.
ATTP
Re: SH warming. Perhaps suppression or cessation of the AMOC, which cools the NH and forces the SH ocean to warm. AMOC shutdowns are thought to be triggered by freshwater flux at high NH latitude – typically the result of increased insolation melting ice. Just how soon the increasing high latitude TSI from orbital forcing triggers enough melt to impair the formation North Atlantic deep water is interesting.
BBD,
Thanks, so it’s still associated with an orbital trigger?
Ah. This is what I was looking for:
WAIS Divide Project members (2013) Onset of deglacial warming in West Antarctica driven by local orbital forcing:
ATTP,
Concerning the strength of internal variability, it’s a plausible alternative that variability on the decadal scale is of the order of the warming to around 1940 or of what’s needed to have the series of flatter and steeper phases in temperature trends over the period 1950-2014. Various forcing have influenced the temperatures, but their roles are not understood well enough to make such ideas implausible.
On longer timescales there are forcing based explanations for LIA, but they are far from definitive. The earlier history of MWP etc. is still less known, but historical data can certainly not show that the variability has been insignificant. Extending to thousands of years, new uncertainties pop out. Several components of the Earth system have or at least may have memory (or persistence) at a level that might support significant variability.
It’s absolutely wrong to conclude that such variability is not likely to exist on the basis that there isn’t strong evidence. To make such statement evidence on the absence is needed, not only absence of unambiguous evidence for the variability.
When the question is restricted to the warming since 1960s the situation is totally different. It’s different, because such warming from added CO2 was predicted and the because the mechanism from GHGs is known. Here we compare something rather specific to the open alternative of something else. Based on the principles of Bayesian inference, the observation gives strong support for the specific and foreseen explanation, i.e. for the AGW.
In all fairness, I would suggest that SoD abides by skepticism by asking questions and trying to figure out the answers for and, more importantly, by himself.
To rely on authorities (say to orient where one should look or to confirm that one is not reinventing the wheel) does not undermine self-reliance as long as anyone thinks trough and works things out by oneself.
A foolish consistency is the hobgoblin of little minds.
Pekka,
I think maybe you read too much into what I say and don’t actually interpret it as I intended. For example, what I said was
I would argue that that is not concluding that it is not likely to exist. It is suggesting that there is little evidence to suggest that it does. A subtlety maybe, but still the main point. So, yes, maybe it does exist. Yes, maybe there are scenarios where internal variability could play a major role in the 1910-1940s warming, the LIA, the MWP, …. However, I’ve seen little evidence (other than the possibility that it’s plausible) to support this possibility. On the other hand, there appears to be much more evidence that most of those eras were forced. So, you seem to be arguing that it is plausible. Fine, I agree. However, the evidence in support of this seems weak at the moment. Hence, little evidence to suggest that it is likely to play a major role in the timescales that are of interest to us today. This may change, but that possibility doesn’t mean that I should suddenly consider it likely (which – I realise – is not what you’re suggesting).
ATTP,
A word can make a difference. I would have taken more positively what you wrote, if you had left the word can out from .. internal variability can play a major role ...
With the word can the claim becomes too strong in my view. It’s a much stronger implied claim that it could not than that it did not. Including the word can requires more evidence from your side.
Science deals in probabilities. It is more likely that significant paleoclimate variability is externally forced rather than self-propelled. In the case of deglaciations ~2.7Ma – present it is very likely that orbital forcing is the trigger mechanism. Self-propelling deglaciation that just happens to occur at high obliquity every 41ka for a million years and then flips to multiples of 41ka for ~800ka is implausible. Suggesting that there are complex, unlikely solutions to problems which are unlikely actually to exist isn’t going to be productive except of confusion and uncertainty. There’s already plenty enough of both around as it is.
Pekka,
I think one might call that playing semantics. Also, leaving out the word can would seem to make the sentence nonsensical. Sometimes I just put words in that allow me to join one part of a sentence to the next part of the sentence.
ATTP,
That particular word really makes a difference for me, not a small difference but a really big one. Without can it’s a statement about one actual history, with can about all possible histories, and that’s really a big change.
Pekka,
Firstly, the can referred to the future, not the past (i.e., “seems unlikely that it can play a big role in the coming decades”, was what I was getting at – could have used will, instead of can). Secondly, once we’ve had this discussion and agreed that maybe a statement of mine was too strong, or that you’ve over-interpreted it, or maybe we just disagree, maybe we should just stop. I can’t travel in time, so can’t change what I wrote. I could go back and edit the comment, but then the subsequent discussion wouldn’t make sense.
So, if the word can makes a big difference to you, what it’s doing is making a big difference to your interpretation of what I said. It doesn’t make any difference to what I was intending to say. Also, an argument/discussion about the word can is amazingly tedious – I am kind of smiling because I can’t quite believe we’re doing this.
Let me clue you in on something. When I write something, it’s written. If someone want to discuss it, disagree with it, dispute it, that’s fine. Once we’ve reached some kind of agreement (“yes, I agree that what I said was too strong”, or “I see your point, but I didn’t mean it the way you think I did”) I really do fail to see the point in continuing. Why carry on pointing out that you don’t agree with your interpretation of what I said, if I’ve already agreed with your general view? I don’t see the logic in that.
No, not a campy comment, SoD. Whenever I say wow I give Judy appropriate credit. I should take more care to ensure people realize it’s not aimed at them. The glass half empty bit had what I hope is the obvious meaning. As a general matter, I do appreciate your careful approach to things.
Pekka: Rather small coverage issues might distort the results, when the changes vary so much from one part of the oceans to another, and there are still issues even with the present ARGO, and more before (including the first years of ARGO).
That remark would best be accompanied by some references to publications describing “issues with the present ARGO.” that are relevant to the doubt Pekka implies is something to worry over. One implication of Pekka’s remark is that the folks employing ARGO data to report on total ocean heat content are making fairly elementary errors of oversight, a rather strong charge to casually drop without further explanation. We’ve enough of that sort of thing going on already, more is not helpful. Pekka, can you please elaborate?
Thanks for turning that up, BBD.
Me: “SoD, climate change is very small relative to the spatial variability and diurnal and annual cycle. Thus for most processes on small spatio-temporal scales, the response is likely nearly linear, even if those processes themselves are nonlinear.”
SoD: “Thus?”
Thus, I found your argument (reprinted below) from the Lorenz model a bit too general and cloud and water vapor not the best example:
SoD: “But for now.. a non-linear system can exhibit variability that is “unexpected” from the drivers that are forcing it. The simple Lorenz 1963 example – shown in Natural Variability and Chaos – Two – Lorenz 1963 is probably the best known.
Well-known physics professor Pierrehumbert gives a great example in his 1995 paper. I highlighted this paper in Clouds and Water Vapor – Part Five – Back of the envelope calcs from Pierrehumbert.”
SoD: “Can you talk me through the second sentence and how it is obvious from the first sentence. I can’t see the connection. Perhaps defining “small” and “nearly” would be a help.”
Defining small and nearly would not help much. Like I wrote before this is just a qualitative argument.
The second sentence follows from the first one in that if the change is small it is normally a good approximation to assume a linear relationship. In case of clouds or water vapour a small temperature change will just mean that they form at a somewhat different height, at a somewhat different time of day, in a somewhat different day of the year, at somewhat different location.
The relevant time scales for the ice sheets, vegetation and thermohaline circulation is much longer. That also makes global warming much larger relative to the variability at their relevant time scales. Then it is much easier to get a nonlinear response, like the abrupt changes in the ice sheets seen during ice ages.
This is all purely heuristic. If you do not buy it fine, there is no science to force you to believe it. Such thoughts help one to find areas where one can productively work. I would personally not look at cloud processes if I were interested in nonlinear climate responses or at the question how much the climate sensitivity may change in time and space. Then I would first look at ice sheets, vegetation and the thermohaline circulation.
Victor – that’s standard, dare I say it, engineering models. If the changes are small enough stick with linear approximations. More formally, you are entering the realm of perturbation theory. As a mere mortal that may be beyond my ken. Over to the physicists and mathematicians. http://en.wikipedia.org/wiki/Perturbation_theory
Steve
Glad it was useful. I’m actually happy this came up because I’d lost the reference. Bloody scientists, constantly publishing stuff. Impossible to stay on top of even one pile. 😉
If you are there, Steve, what do you think about the implications of combining WAIS DMP13’s findings with Ferrari et al. (2014)?
WAIS DPM13:
F14 seems to imply that reduced circumantarctic sea ice cover will have significant effects on global ocean circulation and the ventilation of the Southern Ocean and so climate. None of this seems to me to conflict with eg. Shakun et al. (2012) or Abe-Ouche et al. (2014); rather it simply expands the picture of possible mechanisms by which orbital forcing can trigger a deglaciation.
I don’t want to get too embroiled in semantics, but I’ll try to give my perspective on whether internal variability can explain more than a small part of the post-1950 warming (or the 1975-2000 warming). If there is 100 percent certainty that it did not, then it would be wrong to say it “can”. If it’s somewhere in the realm of 95 percent certainty it did not, then theoretically “can” is still consistent with the evidence. The latter is my view – it’s consistent with the evidence, at about the five percent level.
The evidence against a major role for internal variability is very strong, but it’s not absolute. GCM simulations appear to exclude it, but they may have missed important factors. In addition,climate sensitivity to GHG forcing would have had to have been much less than estimated in order to assign a less than dominant role to it. More important for me than either of the above, the substantial increase in OHC during those intervals is incompatible with any known or observed mechanism that would permit internal variability to concurrently warm both the surface and cause heat to flow into the ocean over multidecadal intervals, all the while creating a geographical pattern of warming that appeared to match that predicted for GHG forcing. As I’ve mentioned above and elsewhere, energy budget constraints appear to exclude that, and I would again recommend the Isaac Held link for a quantitative discussion of the relevant principles. The paper by Brown et al is very much in line with these principles, and is useful in the sense that it’s a model result that doesn’t contradict the dominance of GHG forcing. As Pekka has noted, that in itself is not a proof.
Nevertheless, based on the principle that “anything is possible”, I expect one could come up with a scenario that allowed internal variability to create most of the observed warming and the pattern of warming, while circumventing the difficulty of adding heat to both the air and the ocean below the surface at the same time. What could it be? I have no idea. I suppose that if all clouds in the tropical Pacific, for reasons known only to them, decided in 1975 to move to the Arctic and stay there for 25 years, that would have done it by increasing SW absorption in the tropics and LW GHG effects in the Arctic. That obviously wouldn’t have matched observations, but it’s conceivable that observations might have been matched by a process that was subtle, unsuspected, undetected, of unknown mechanism, and absent from unforced model runs over hundreds of years. How likely is I suppose a matter of judgment. I would say maybe about five percent is not unreasonable, although it’s probably too charitable.
Thanks for that link, BBD. I’ll have to read it more carefully, but it sounds good. Broadly, it would make sense that increased SH insolation would at the very start act to melt sea ice and accelerate the westerlies, with CO2 outgassing and ocean current changes commencing immediately.
Doug,
What I had mainly in mind are the remaining areas that are not covered well. When the rate of change of the temperature varies sharply based on the location and also on time as it seems to do in several places, areas of weaker coverage may be more significant than the most straightforward error estimates tell. That there are still some such issues, was only a side remark, but coverage was significantly worse still during the first few ARGO years. Therefore the period of highest quality data is really short.
willard:
“This means pointing out that “[m]any other people with a publication record also ask these questions” could be taken as an argument from authority.”
Or it could be simply pointing out that someone who invoked the argument from authority needs to present evidence for their claim, seeing as there is evidence against their position.
Steve Bloom said: “Despite all of SoD’s “unknowns,” when scientists first started grappling with this problem 40 years ago it was entirely obvious that the timing of the post-MPT glacial cycles was keyed to eccentricity.”
I took that as, “..everyone in climate science sees this. Obviously people questioning this basic fact are not paying attention.” Therefore, I put forward evidence against.
Later you said:
“In all fairness, I would suggest that SoD abides by skepticism by asking questions and trying to figure out the answers for and, more importantly, by himself.
To rely on authorities (say to orient where one should look or to confirm that one is not reinventing the wheel) does not undermine self-reliance as long as anyone thinks trough and works things out by oneself.”
This is what I do. I read textbooks and papers by climate scientists. I assume that textbook physics is correct. I assume that people producing papers understand their field. I learn from these experts.
But it seems that questioning the strength of evidence of what has been received into the world of papers and is “widely believed” is not popular here.
Ah. Thank you Pekka.
Here are some interesting articles exploring or touching on the general matter of Argo spatial coverage effects on S/N, assimilation into synthetic datasets, etc.:
Click to access osd-8-999-2011.pdf
http://onlinelibrary.wiley.com/doi/10.1029/2010JC006910/full
http://journals.ametsoc.org/doi/pdf/10.1175/1520-0426%282004%29021%3C1598%3ADRFAAF%3E2.0.CO%3B2
http://onlinelibrary.wiley.com/doi/10.1029/2005GL025552/full
“Therefore, I put forward evidence against.”
You pointed to Wunsch and Hubers and to Wolff et al. I looked at the most recent paper from Huybers and quoted his reference to eccentricity. Wolff et al. don’t discuss eccentricity directly, referring more generally to “variations in earth’s orbit,” but note that the first thing they do is cite to Hays et al. (1976), which says in part:
I strongly suspect if I looked at the two earlier Wunsch papers I’d find similar.
There are too many ~100 ky cycles in a row for anyone to think that glaciations aren’t in some manner tied to eccentricity. Yep, it’s by no means the only thing involved, but that wasn’t the point I was making.
So why are you trying to argue about the bleedin’ obvious?
Because the scientific evidence is inconclusive and obviously incomplete as to the exact mechanisms. But this is why I ranted a bit about the role of probability in the evolution of scientific understanding. The scientific consensus is, as Pekka points out, capable of incorporating uncertainty without becoming dysfunctional.
Of course, BBD, but that’s not what he was arguing about.
[Mod: This comment has been removed by the moderator]
Peter seems to be working on other things these days, but I’ll try emailing him to ask what he thinks.
> This is what I do. I read textbooks and papers by climate scientists. I assume that textbook physics is correct. I assume that people producing papers understand their field. I learn from these experts.
Indeed, there is a way to appeal to authorities that can be considered valid. [This was my] main point, i.e. to show that what people say can be used as evidence. At the very least, it provides evidence of what they say. At best, it makes us co-create social products, like scientific knowledge.
Nullius non verba always seemed to me a bit farfetched.
***
Come to think of it, I should clarify this other appeal to authority by SoD in the quote I underlined: “as a skeptic”. To claim that one is a skeptic can be considered as an argument that adds credential to what one does. This claim is valid if and only if it leads to a description that can be verified. In SoD’s case, it’s quite clear that he asks questions and tries to figure out the answers for and by himself.
Whatever the validity of that claim, that SoD relied on this self-avowal can be understood as a way to confirm that his endeavour was justified. This confirmation seems to have been needed to parry cross-checks. That some cross-checks were felt can be seen in that impression:
> [I]t seems that questioning the strength of evidence of what has been received into the world of papers and is “widely believed” is not popular here.
I agree, at least insofar as it’s a more or less generalized phenomena. See for instance this thread at our beloved Bishop’s:
http://bishophill.squarespace.com/blog/2014/8/24/gcms-and-public-policy.html
I would venture that the blog alignment is recognized by how well questioning is received. We should not lump every commenter in the same category. Just like other recognition task, it takes time to recognize the different voices from the commenters of a place.
Pekka:
Actually, that does seem implausible to me. Whatever the cause of the 1910-1940 warming, some of it is known to be forced because we know there were positive anthropogenic and natural forcings in that period.
It is reasonable to assume that the ratio of the forced response between 1910-1940 and 1950-2010 is equal to the ratio of forcings between the same periods. Further, internal variability from 1950-2010 is very close to zero, both because of the length of time relative to the duration of short term variability such as ENSO, and because the period closely approximates to the duration of proposed long term variability. From this we can determine a plausible limit of unforced variability in 1910-1940, which is about a two thirds of the total.
We can also take an alternative approach of using proxies for proposed sources of variability into the past and comparing their past magnitudes with the 1910-1940 variation. The difference becomes the estimate of unforced variation in the early twentieth century. If the proxy is variable in magnitude in prior centuries, the mean amplitude plus 2 SD becomes the upper limit on presumed natural variability. With the AMO, and using the CET as a proxy, this again leads to the position that most of the 1910-1940 temperature increase was forced, though it is plausible (again) that about natural variability accounts for about a third.
This looks stunningly like a hypothesis of the gaps to me.
What we can really say about such uncertainties is that is is possibly due to unforced variability, or due to measurement error, or due to inaccurate estimates of forcings, or due to non-linear response to forcings, or due to …. Until, and unless you can complete that list and assign relative probabilities to the alternatives on the list, the uncertainties in question are not evidence for any of the hypotheses in question. They are just uncertainties in testing the hypothesis we know is on the table, ie, that changes in forcing drives major changes in global temperatures.
[Mod: This comment has been removed by the moderator]
I linked to the paper above. 2013 is correct.
Will SoD be as happily aligned with the AR5 in 5 more years?
Damn. I just lost a long comment that described step-by-step what I did to end up with that question. Here it is:
Where did SoD say that A13 made a claim about the initiation of the glacial termination, and how does this claim embrace everything that SoD said about the subject?
Something’s amiss.
There are some other factors that needs to be considered when deciding how much weighting to give to 1910-1940 surface temperature data. Spending some time at GISTEMP looking at the zonal and map diagrams rather than just the headline trend graphs is interesting.
1. Station Coverage. At the start of this period there was virtually no met’ station coverage in the Arctic – this grew during this period particularly as the Soviet Union stated establishing weather stations through Siberia. And there was no station coverage at all in Antarctica till the late 50’s. So we can’t rule out coverage bias issues as a contributing factor to the early record.
2. Arctic warming. A significant part of the warming seems to have been in the Arctic, perhaps particularly the Atlantic sector. Anecdotal support for this comes from the fact that the Dust-Bowl years in the US coincided with this period (I know that not all the Dustbowl was just climate – also agricultural practices).
Put these two together and you potentially have a local rather than global change, perhaps being overrepresented in the trends because station coverage was increasing in that region just as the local warming occurred. And if it is localised then we need to look for more local explanations. And the big one that leaps out would be any fluctuation in Atlantic/Arctic ocean currents – any variability in the AMOC is going to impact the NH particularly.
So 1910-1940 can tell us some things but may not be as reliable as we might thhink.
Ahh… the calm noise of a technical discussion.
Weren’t we talking about something similar recently? How far can models accurately project forward and given the Holocene conundrum, how far back can we go?
There’s the error. Later he semi-retracted, but persisted:
As I noted, neither Abe-Ouchi et al. or I had made such a claim.
As to whether it embraced everything, I wouldn’t know. It certainly seems to have derailed the discussion.
Tom,
1) I don’t present any hypothesis. I only point to the fact that we have observed variability, and that variability has not been explained even nearly fully. Thus it’s plausible to assume that the magnitude of the variability is indicative of the magnitude of internal variability. It’s not a proof of the strength, but it’s a plausible reference for setting the scale.
2) My reference to the period to the period 1950-2014 was meant to refer to what’s left, when a smooth AGW trend is removed.
One more point is that ENSO is internal variability and that variability in the relative weights of El Nino and La Nina is longer term variability.
Glenn (and others with the required technical skill),
In principle the effect of this change in global coverage could be investigated by only looking at the stations with 100+-year records, right? I know it will not be perfect,
Steve,
I cannot accept the statement that we have so many 100 ky cycles that that would provide strong evidence for the role of the eccentricity. We have only a few rather irregular ones and we have different cycles before.
Abe-Ouchi provides an alternative explanation for the 100 ky period as based on internal dynamics of the Earth system.
BBD has understood me correctly:
It would be really nice if the disagreements about the climate science itself could be discussed separately without the perceived need to think every time, how the claims affect policy conclusions.
The discussion gets heated in other fields of science as well, but in most cases not in the same way it gets, when climate science is discussed. I know that my wish is unrealistic for the internet as whole, but all sites need not be as badly influenced by the issue.
SoD,
Not sure that’s an entirely fair statement, but – if it’s what you think – fine.
On a related note, I get the impression that you and Pekka both see yourselves as being amongst the only true skeptics. I’ll admit that I’m impressed by what you do on your blog. A great resource and you clearly understand this topic well and put a lot of effort into reading, understanding and analysing papers. Similarly, Pekka is extremely knowledgeable, well-read and has a good understanding of this topic. What would help, though, is if you both applied a similar level of skepticism to your interpretations of what other people say.
…and Then There’s Physics said:
“Not sure that’s an entirely fair statement, but – if it’s what you think – fine.”
If you visited my blog and challenged ideas but you were accused of misrepresentation, smokescreen, having other agendas you might wonder whether you were welcome.
It’s not possible to have a useful discussion in such an environment. I’m sure you would feel the same.
I understand that I am not welcome here. It’s your blog.
“I get the impression that you and Pekka both see yourselves as being amongst the only true skeptics”
I can’t speak for Pekka. What makes you say that about me? What have I said?
I think I said that I believe it is ok and good to ask questions. To challenge ideas. Many people do this. Many climate scientists do this, and I know because I read their papers. So no, I don’t believe I am in some elite group.
I feel I need to start explaining the “I am a skeptic” point because it seems that as soon as I am perceived to be “questioning stuff that people believe”, one of the regulars is accusing me of poor motives. So I explain, no, I think asking questions is good. And now, that means I think I am “a little bit special”?
It’s not an environment where questions can survive. You asked the wrong question, clearly you are doing it for poor motives. You explained yourself, clearly you think you are special.
What’s the right approach here on this blog? Don’t question people? Don’t question ideas? When people tell you are wrong, accept it? You should hand out a form in advance. Legitimate questions list. Bad motives questions list.
I have one on my blog – “Basic science is accepted”. So people know their ideas are just going to be removed if it they are starting with textbook physics is wrong. After that, it’s ok. Of course, I have other rules, constant repetition, long essays.. you can read the list.
“What would help, though, is if you both applied a similar level of skepticism to your interpretations of what other people say.”
So there’s an implied criticism. Why not be specific? I’m not skeptical enough of … what exactly?
I’m fine with being challenged on ideas and what I think. Poor science should be challenged. But how can I respond to this criticism? I don’t know what it is. So why not, now you have hinted at it, get it out in the open?
SoD,
You are more than welcome. I am genuinely impressed by what you do on your blog and hold your views in high regard – even if I don’t always agree with them.
Maybe I’m missing these, but I didn’t think that such accusations had been made. Haven’t read all the comments though.
This was based your apparent suggestion that people here aren’t being sufficiently skeptical. You appear to be being quite critical of me and of others. Maybe I’m misunderstanding you, so it was just an observation.
Fair enough, it was just an observation. Apologies.
Again, apologies, but sometimes when people explain themselves it can come across as a suggestion that somehow they are behaving in a way superior to others.
I don’t follow this. Are you accusing me of doing this, or not?
Again, just an observation. I was just suggesting that some of the discussion on this post has ended up appearing to be critical of the behaviour of others. You obviously feel that you’ve been accused of misrepresenting things while I think there has been a suggesting by yourself (and maybe Pekka) that others aren’t being sufficiently skeptical. Maybe we should all stop being defensive.
It’s unfortunate that this discussion appears to have degenerated. My previous comment wasn’t intended to be insulting. It was intended as a bit of a push-back against some of the apparent criticism from yourself (and maybe from Pekka). As I’ve already said, I hold you both in high regard and think your understanding of this topic is excellent. You certainly know much more than me about some (maybe all) aspects of this. It’s just a pity that these discussions can end up in this form rather than sticking with discussing the science. Apologies if my previous comment seemed a bit too personally critical.
ScienceOfDoom,
You are very welcome here (although it’s not my blog so I might be stepping over the line a little bit by saying this). However I do think this is a bit unfair:
It is perfectly acceptable to challenge the evidence (within reason – comments from people denying the greenhouse effect might get moderated but plenty of these still make it through). However you have to expect that your own comments will also be challenged.
The main requirement for commenting here is that commenters treat each other with respect. If you feel you haven’t been treated with respect then as a moderator I apologise.
There are many papers showing that changes in ocean heat transport can change the energy budget substantially starting at least as long ago as the 1970s and extending to recently. I convinced John N-G to link a few at the Climate Change National Forum when Andreas was making an argument about internal variability and OHC there. You should take a look. This is not a new concept nor is it a dying one.
capemall,
Thanks, but my understanding – which may be wrong – was that most previous work had considered inter-annual variations and that this was one of the first to consider periods of a decade.
Don’t know what makes you think people here think it’s dying.
Have a look at my first 10 comments. I think I’m questioning. I’m criticizing. I’m putting forward an argument.
I do accept, and acknowledge, that my questions can seem confrontational sometimes. I often wish I was like some of the people I know who have a very different, gentler approach. But I’m not.
And – how to crystallize the issue? How to ascertain the key points that decide on the difference between one idea and the other. The way I know – asking questions, and then asking more questions. And then relentlessly following up.
Picking up an idea and turning it around. It’s my way and some people don’t like it.
After question 10 it is suggested:
“Wow (tm Judy Curry), SoD, your cup sure is half empty on this one.”
I am aware of the views of Judith Curry held by regulars on this blog, because I often read this blog without commenting.
My response, which began with trying to explain my ideas, concluded with “I call myself a skeptic. This means asking questions and not accepting the argument from authority.”
So comment 11, after being apparently identified with the much-vilified Judith Curry, is trying to explain my approach to the subject. First 10 comment I asked questions.
Then we have willard (@nevaudit) comment of August 27, 2014 at 2:29 pm which I can’t even understand but it doesn’t seem nice.
Steve Bloom, August 27, 2014 at 11:23 pm:
“[Mod: refers to a deleted comment]”
willard (@nevaudit), August 28, 2014 at 1:32 am
“[Mod: refers to a deleted comment]”
None of this seems like a discussion about climate science. Clearly I am in the wrong blog. Smokescreen? Lovely approach.
I actually think these people haven’t understood the subject they are arguing about but I don’t accuse them of falsifying their motives. I just assume they haven’t understood the subject matter.
Of course, it could be me who has missed the point.
Crystalizing the issue is the key to figuring out where the difference of ideas lie. That’s my idea.
Alternatively, just insult the other party. That’s the approach here.
In one of my first encounters here, one of your regular commenters accused me of misrepresentation and lying. I was lucky to have another of your regulars email me privately and say that they believed the suggestions were completely unjustified. This was very helpful as it made me consider returning.
The blog title is nice. Living up to it is difficult. If you want people not to be insulted for different ideas you need to comment that you don’t agree with the insults. Otherwise visitors like me will assume this is part of the treatment for asking questions.
Just to flog the dead horse a little bit further:
“You obviously feel that you’ve been accused of misrepresenting things while I think there has been a suggesting by yourself (and maybe Pekka) that others aren’t being sufficiently skeptical”
There is an important difference between ideas being questioned and integrity being questioned.
Well, I thought there was. We can all question each other’s ideas, assuming we are interested in science.
When you question someone’s integrity it is a different ballgame.
Whose integrity have I questioned? If this is your blog, doesn’t it bother you when visitors arrive and your regular commenters accuse them of misrepresentation? Do you agree with it? You don’t care? It’s the same as questioning ideas? It’s all fair game.
“Smokescreen”. Apply the statements to your own comments and think how you would react.
Can you explain your blog policy in this area?
An example, by way of explaining, if it should be necessary.. A long time ago I commented for a while at one blog, to remain nameless, well-known for its awful understanding of basic physics, and I happily endured many insults about my “lack of knowledge” of heat transfer, I suggested many criticisms (insults?) myself of regular participants’ understanding of heat transfer basics. But finally the blog owner accused me of dishonesty and I never returned. I assumed this was his intent. (Well, to finish the story, finally he kind of apologized, so I did make at least one guest appearance more recently, mostly to return to my theme of mocking the appalling approach to science – but still, not to question people’s integrity).
SoD,
I normally would and – like you on your blog – I comment below the line quite a lot. Rachel is also a very good and fair moderator. Lately, however, we’ve both been busy and the blog has been quite quiet on the confrontational side for a while, so we’ve relaxed a bit. Also, some of the rhetoric in this debate can be so awful, that I think I’ve recalibrated what I regard as acceptable or not.
I’ll make a comment about this though,
As I think BBD was trying to point out, almost all of science is associated with developing an understanding of some topic and how likely it is that that understanding is “right”. In some cases we can be virtually sure, but in others, not so much. Given this, there are certainly situations where well-informed people can disagree and both have reasonable arguments for their positions.
The Milankovitch cycles seems like an example of this. We don’t have a well-accepted mechanism, but most of the evidence appears to suggest that orbital forcings might play a role. So, someone like me may think that it’s probably triggered in some way by changes in the orbital forcings. Someone else may think that we don’t yet know enough to be sure. Both may be reasonable interpretations and I wouldn’t think poorly of someone for thinking differently to me.
The other issue is the relevance of all of this to today. So, there is the purely scientific interest, in which case we can simply discuss the different lines of evidence and the strength of these lines of evidence. On the other hand, the reason this is an interesting topic is because of what we’re undergoing today. Could internal variability have played a significant role in the overall warming over the 20th century. My understanding is that this is unlikely. It may well have contributed significantly during certain periods, but is unlikely to have contributed significantly to the warming since 1950 and – as I understand it – the overall warming during the 20th century is also likely to be predominantly due to changes in external forcings, rather than to internal variability.
Similarly, could it play a significant role over the coming century? Again, my understanding is that this is unlikely. It can certainly influence how we might warm over decade timescales, but will probably average out over the longer time periods. So, the possibility that internal variability could have played a bigger role in some of our past climate changes than maybe we currently realise could have implications for our understanding of climate change today, and in a sense it certainly does. However, I would argue that until we have evidence that it probably did play a bigger role than maybe we currently think, assuming that it’s role today will probably be small, seems reasonable. You may, of course, disagree, and I’m certainly not suggesting that we should now stop considering the role it may play – I wouldn’t have written this post if I didn’t think it was an interesting and relevant topic.
I’ve written this quite quickly, so maybe I haven’t quite added all the caveats and uncertainties that I should have, but hopefully you get the gist of what I’m trying to say.
SoD,
I’ve apologised. I don’t really know what else to do or say. I’ve never claimed to know what I’m doing or how to run a blog. I’m trying to do the best that I can. If it’s not good enough for you, that’s fine. I think your understanding of this topic is excellent and I’d be disappointed if you felt you couldn’t comment here, but I can’t go back in time and change what happened and I can’t claim that I’ll always get it right in the future. Also, as much as I’m happy to acknowledge my failings and apologise when things don’t go as they should have done, even I have my limits.
Thankyou, Rachel M.
…and Then There’s Physics – from August 28, 2014 at 10:31 am:
Thanks, that is appreciated.
ScienceOfDoom,
It’s a difficult balance knowing what comments to moderate and what to allow. I think most people prefer very little editing of comments. My first instinct is that the comments in question that refer to smokescreens are not insults. But perhaps there’s an implied deception in the use of that word. So I’ve gone ahead and removed those comments. I’ll have to edit the quotes out of your own as well.
> What have I said?
“I call myself a skeptic. This means asking questions and not accepting the argument from authority.”
Labeling oneself has an impact on others and oftentimes feeds personalization.
This is why I replied “and I’m a ninja,” a meme that has currency on the Internet.
***
> Then we have willard (@nevaudit) comment of August 27, 2014 at 2:29 pm which I can’t even understand but it doesn’t seem nice.
Not understanding something is a good time to ask questions. It may not be the best of times to start network analyzing. See the remark about those who comment [on Judy].
In that comment, I was reminding that there are valid uses of argument from authority. This is a philosophical point that matters to me.
Whether it’s nice or not is the lesser of my concerns.
***
> Whose integrity have I questioned?
SteveB’s, mostly. Wondering [about] the blog’s tag line may question AT’s. With the network analysis, probably me [too]. And that’s notwithstanding all those who rely on authorities, which I believe comprise [everyone], including those who can be grouped with the “august” group of AR5 writers.
Just in case “network analysis” is unclear, it refers to “after being apparently identified with the much-vilified Judith Curry”. This identification seems to have been read into SteveB’s “wow” comment. But then I get lumped in that analysis with “Then we have willard” and “I actually think these people haven’t understood the subject they are arguing about”.
I’m not sure “these people” argue about the same “subject” at all. My understanding of appeals to authority and labelling can be substantiated. For instance, here’s a starting point on the use of expertise in argumentation:
Click to access 89reasoned.pdf
There’s a whole research field on labelling is social psychology. There’s even a theory about labeling:
http://en.wikipedia.org/wiki/Labeling_theory
***
Just in case SoD gets the impression that I’m here to root for SteveB, here’s something for his network analysis:
Had I had not lost that long comment where I was showing that the “smokescreen” comment was itself a smokescreen, we might have reached a different state.
Pekka: It would be really nice if the disagreements about the climate science itself could be discussed separately without the perceived need to think every time, how the claims affect policy conclusions.
Pekka’s thought pretty much mirrors mine as I’ve watched this discussion unfold. Thanks to being exposed to some many obtuse correspondents, a lot of us (myself included) suffer at least a little bit from the intellectual equivalent of an irritated and overly-sensitive immune system, to the point sometimes of resembling autoimmune disorder with some of what that implies in terms of rejection of healthy, functional organs etc.
Rachel M: This is a textbook geek argument. Usually it ends with everyone parting ways in violent agreement, or confused. 🙂
The difference between this and denier trolling, is the lack of malice, and earnest responses between rants. People here are trying to seek consensus however haphazardly. One of the biggest hurdles for all folks here is that they all come from different backgrounds and frankly have been simmering in different perspectives. Trying to reconstruct the exact bits in a way for each other to understand can be tedious at times. (I used to run an R&D team.)
I wonder how it fits with Willard’s studying of correct technical argumentation. 🙂
> Thanks to being exposed to some many obtuse correspondent […]
I thought I was alone on that category.
Who else but me do you have in mind, Doug?
(In case people wonder what the I’m doing right now, I’m illustrating how easy it is to become defensive, and underlining a figure of speech that shows that civility does not reduce to a calm and polite tone.)
Folks: Being adults and all, you could start a new thread. As I said, there is no malice. This is just arguing.
> This is just arguing.
Perhaps we could move on to contradiction, or abuse.
Per Willard’s remark and assuming whole-body x-radiation treatment is out of the question, what’s the equivalent of corticostearoids for conversation? Humor is my drug of choice. 🙂
This is an interesting discussion. Victor May place more faith in linear perturbation theory than is justified. In any case, for GCM’s to actually usefully apply it is practically impossible. One needs to compute wigs values of a very large and ill-conditioned matrix. For the Lorentz system useful information can be computed, bolstering the idea that simple models can be very useful.
Spell checker strikes back. Eigenvalues not “wigs values.”
And if so, you know just how to fit in.
It’s much worse than just an eigenvalue problem for a time dependent system. The adjoint requires a backward integration in time and for chaotic systems is infinite and not computable. The adjoint represents the linearized sensitivities of the result with respect to changes in say initial conditions, parameters etc
David
I know you spotted this immediately, but I just have to repeat it:
You shouldn’t correct your stuff. It’s vastly superior in draft.
🙂
I thought it was Whig values.
Steve Bloom: I have kids… This is no different.
After a while they go away in a huff. The trick with those arguments is that you need to get past “He said, she said, he said, she said, he said, she said, he said, she said, he said, she said, he said, she said, he said, she said, he said, she said.”
There is no longer a purpose in determining the start of an argument. Just come back and start a fresh.
Anders; Repost the article… and ask the SOD, Pekka, others, and possibly Steve (:-)), to come back. Maybe go at this a bit more slowly, and a bit more precisely. Willard is of course not invited because he keeps Whigging out.
I just read the Dot Earth post on ocean variability. All the comments presented there are interesting and make sense (while some contradict each other a bit). Both Revkin and all the scientists interviewed made good points and largely points that I agree on.
(Yes, I got the link from Climate Etc before I found it otherwise. I follow regularly NYTimes reading or at least skimming all posts of Revkin – and of Krugman – but not so tightly that I would have picked the post that way.)
Pekka,
I haven’t read it closely, but the Do Earth post does say,
The latter part of that is not correct, as this came from the press release and was based on an earlier paper.
ATTP,
The present paper does not state that directly, but it seems to add support for the statement – at least on semiquantitative level, I’m not sure whether fully quantitatively.
Mr Young, nice to hear someone who knows something about numerical fluid dynamics and chaos theory. Could you explain how that is relevant for estimating climate sensitivity? Models are simplifications and by definition wrong. Thus you do not only have to show that something is wrong, but that that matters. Feel free to use your skills to develop better numerical schemes for atmospheric and oceanic models. If your model would have a different climate sensitivity because of the numerics, i will write a post titled: “Oh no, now I need to eat my words!” If you do not find any differences, the numerical weather prediction guys may still be happy with your work.
Is it a case of http://imgs.xkcd.com/comics/physicists.png or http://derp.co.uk/xkcd/comics/purity.png or both, Mr Young?
Victor, I merely assert that linearity is a bad assumption for these systems and in any case, even linear stability is uncomputable for time dependent systems whose long time behavior is sought and that are chaotic if they are at all large. That is the beauty of simple steady state models if they are properly constrained by data their behavior is at least computable. I have no goal to improve numerical weather prediction which seems to me to be already near the limit of computability. Just running a tremendously complex simulation (that no one person completely understands) over and over again is in my view not a way to advance any field.
In this regard SoD is right in that our understanding must be incomplete for these systems. I understand his annoyance at people who try to make everything seem simpler than it is and more settled than it is.
> That is the beauty of simple steady state models if they are properly constrained by data their behavior is at least computable.
Now I’m interested. Got a link?
I hope you do not try to make everything seem simpler than it is, David.
So do I. We end up with EBMs and you know where that leads.
And as for choosing between Whigs and Tories…. well, thank god for the oblivion of drink.
Pingback: I think I need another break! | …and Then There's Physics
jsam: Funny. Oil and gas gets whacky ideas all the time.
And R&D Manager at one company asked me why we couldn’t use GPS down hole. I said, “Ever use a cell phone in a tunnel?” The Manager nodded sagely, and said, “Ohhhh… I seee.”
AoM,
I was once on a trip where we were using a GPS to navigate a vehicle. The group leader was getting more and more frustrated because it didn’t seem to be working. We eventually pointed out that the antenna was lying on the back seat.
I see that the Rohling group has a new paper providing some pretty strong support for Abe-Ouchi et al. (2013):
37 is Raymo (1997) and 38 is Abe-Ouchi et al. (2013).
I linked it over at SoD too. We’ll see what he makes of it.
SB –
The only problem being the rate of warming, which is very different now. Plus possible distribution differences – at a guess, there is a difference between GHG-driven warming, with polar amplification and warmer nights, and Milankovitch-driven warming.
TBC, I wasn’t trying to make a reference to current warming, Andrew, nor does either paper. The kerfuffle (above) was all about whether Abe-Ouchi et al (2013) had found the cause of the Northern Hemisphere post-Mid-Pleistocene Transition deglaciations.
The lesson for the present seems to be that ice sheets can respond to forcing quickly, although Grant et al. (2014) (the Rohling group with co-authors) find that the less the southerly extent of the NH ice sheets at the time of initiation of collapse, the slower the process went (presumably because the ice-elevation feedback doesn’t work as well on thinner ice sheets receiving less insolation).
In any case, as you say our present concern is different, the relevant paleo-analog being the rapid Plio-Pleistocene obliquity-paced fluctuations of the Antarctic ice sheets described by Naish et al. (2009). The scale of those fluctuations over 40 ky periods means that they will have been fairly fast (as these things go), but the Naish et al. sediment proxy lacks the resolution to provide much of an upward constraint on the process. (FTM, Grant et al. also has a smoothing issue that probably caused them to miss some peaks, although at a finer scale.) I don’t know if it’s possible for the Rohling group to extend their proxies back to pre-MPT times, or even more usefully back to the mid-Pliocene, but hopefully they can.
That said, the Abe-Ouchi et al. paper is somewhat relevant since the present AIS has a lot more volume/thickness than would have been the case during the Plio-Pleistocene, so there’s potential for an ice-elevation feedback on a scale more like what they studied.