There is a new post a Watts Up With That (WUWT) called a a sea change for climate science. It’s by David Stockwell who is, apparently, an Adjunct Researcher at the Central Queensland University.
David starts with
Climate models can be seen as encapsulating the dominant theory, even though they are composed of many different theories regarding land, the ocean and atmosphere. …… Lets agree, for the purpose of argument, that the dominant AGW paradigm is of global temperature’s high sensitivity to CO2 doubling, resulting in an increase of around 3oC, which appears to be about the central estimate of the climate models.
Well, I disagree that climate models encapsulate the dominant theory. If he’s meaning anthropogenic global warming (AGW) then much of the theory associated with that is basic physics. The influence of increased greenhouse gases in the atmosphere. The feedback associated with increased water vapour. The theory doesn’t rely on climate models. Climate models allow you to understand how the climate system may evolve given certain assumptions about the change in greenhouse gas concentrations, but they don’t encapsulate the dominant theory. Also, 3oC is not a high climate sensitivity. It’s near the middle of the likely range. Furthermore, some of the strongest evidence for climate sensitivity comes from paleo-climatological research and not from climate models.
David goes on to talk about ocean heat content and says
A number of climate bloggers have remarked on the very low rate of ocean heat uptake (here, and here, and here), much lower than predicted by the models (here, here, and here).
The various here’s in the above quote are links. I won’t add them all. What’s slightly odd is that the second set of here’s appear to link to pages that don’t even mention the rate of ocean heat uptake. One of the first set of here’s goes to a post by Lubos Motl. I’ve encountered Lubos Motl on other science blogs and he seems to be regarded by some, and maybe by himself, as a sort of polymath. I can’t comment on his abilities with regards to other science areas, but in this case he is sorely mistaken. He says,
One may also convert the temperature changes to forcing and one gets less than 0.5 watts per squared meter, almost an order of magnitude less than the forcing 3.7 watts per squared meter commonly associated with the CO2 doubling.
What he’s done here is calculate the average rate – over the last 45 years – at which energy has accrued in the oceans and then compared this to the change in adjusted forcing since pre-industrial times. This is similar to a mistake that David Stockwell himself makes and which I discussed earlier. Firstly, the change in radiative forcing would only match the radiative imbalance if temperatures have not changed since pre-industrial times. They have, and so one would not expect them to be even to close to being the same. Furthermore, the average of the radiative imbalance over the past 45 years is quite different to the radiative imbalance today. The comparison that Lubos Motl makes is just completely meaningless.
David continues with
The ‘blow-out’ in the range of likely climate sensitivity can only mean one thing: We are no longer in a period of ‘normal’ science, but entering a period of ‘paradigm shift’.
No, what we seem to have are a number of people who don’t understand the difference between a radiative forcing and a radiative imbalance. The real paradigm shift would be if some people started to at least consider that professional climate scientists are not making the very basic mistakes that some seem to think that they are making.
David then says I personally think that miss-specified models have contributed dismissal of solar influence, and have developed an alternative ‘accumulative’ theory of solar influence and he includes the following figure
I tried to work out precisely what David’s model is, but involves reading papers published on viXra. The hurdle for publishing something on arXiv is not particularly great, so publishing on viXra says something in itself. Furthermore, the only way the system can accumulate energy is if the surface temperature is below equilibrium. You can’t just simply decide to sum the total solar irradiance (TSI). Essentially it seems like complete garbage, but if anyone wants to try and convince me otherwise, feel free to do so (although, given my stricter moderation policy, you’re going to have to do some real physics and not just wave your hands around).
David ends his post with
Climate skeptics don’t want to say we told you so but, well, we told you so. Even though we do not yet have an accepted theory of solar influence, there are 25 unique models in the AR5-sponsored CIMP5 archive, most with a climate sensitivity untenable on observations from the last decade.
Take out Occam’s razor and cull them – deep and hard.
Well, David’s told you so would have more merit if most of what he says didn’t appear to be completely flawed and didn’t illustrate that he doesn’t really understand the basics of climate science. Furthermore, as I point out here, Occam’s razor is really just a guideline for how one would develop a theory/model, and not really a mechanism for determining whether or not a theory/model has merit. I suggested in my earlier post that we should
add a corollary to Occam’s razor which says that if you need to invoke Occam’s razor to make your model seem the most credible, your model immediately becomes the most complicated of all possible models and, therefore, by Occam’s razor is then the least credible of all possible models.
This would seem like a suitable situation in which to apply this corollary.
I have no trouble with saying that the general circulation models encapsulate much of our understanding of the climate system. For computational reasons we cannot model everything as realistically as we would like and make it match our full understanding better.
That probably also explains why the GCM results are so dominant in the IPCC reports. These reports have to synthesis our understanding of the climate system. If you want to do so in one plot and spare everyone having to read a complete book, plotting some GCM runs is a good summary.
Okay, if he just means that GCMs include the fundamentals of AGW, then I would agree. I took it to mean that the theory relied on GCMs which was, admittedly, an interpretation of what I thought he meant. One could interpret it literally, in which case, sure I agree.
Yes, that could also be intended and would be nonsense. There were estimates for the climate sensitivity to CO2 long before computers existed.
I just had a look at the post. Did not visit WUWT for some time. You were worrying some time ago that WUWT would be frustrating. I must admit I find it hilarious. And what makes it even funnier is that they act as if they take it seriously.
Already the first line: “As CO2 climate models falter and even the IPCC backs off its estimates, it just may be that a radical shift in thinking is looming. Wouldn’t it be funny if it was the sun all along?”
That would be funny indeed. 🙂 You cannot code a big enough smiley.
Is there an English word for “fremdschämen”? To be embarrassed because of the behavior of someone who is not embarrassed. They are of the same species after all and most of humanity sees us as one group. There is also a hint of fremdschämen.
And the Stockwell expects that radical shift of thinking, a paradigm shift, due to the hiatus in the surface temperature. How you can talk yourself into the importance of a detail? If the temperature would drop, then we would be on the brink of a paradigm shift. Now we might be on the brink of a small modification of the GCMs. I like variability, it is nice that now more people are interested in such beautiful details.
Victor, I guess I find do WUWT amusing. What I find frustrating is that there are serious commentators (media people) who appear to take it seriously. I do always worry a little that it would be better if I simply ignored WUWT rather than trying to address what is said there. I don’t think there is an English word for fremdschämen, but it does seem like quite an appropriate term.
Victor Venema asks:
Perhaps American English comes close with “facepalm”?
BBD, but I don’t think facepalm includes any sense of empathy 🙂
Stockwell’s “model”, if it makes any sense at all, implies enormously high sensitivity. IPCC AR5 estimates the change in solar forcing relative to 1750 CE to be 0.05 W/m2. If this is responsible for the 0.6K warming since 1950, sensitivity is 12K/W/m2 or 44K per doubling of CO2 (actually much higher as that would assume that temperature is now in equilibrium). Even if IPCC AR5 underestimated solar forcing by an order of magnitude, sensitivity would be at the high end of what is plausible. And magic is needed to stop CO2 forcing.
Stockwell’s manuscript shares viXra with
“New Physics Suggests DARWIN’S Origin of Species is Incomplete, and that Godlike Humanity Will Emerge”
which is about as plausible.
A deft skewering administered by richard telford there.
Richard, that article really exists. I had to check. And the abstract is even better. My emphasis.
New Physics Suggests DARWIN’S Origin of Species is Incomplete, and that Godlike Humanity Will Emerge, by Rodney Bartlett
The basic outline for a different perspective on the Theory of Evolution has been described in this article. It includes many elements that the average person might call nonsense, fantasy or science fiction. I don’t want this hypothesis to be tossed in the nearest rubbish bin, so I’ll explain each element in Part 2 which proposes that the origin of life, and its evolutionary adaptations, cannot be comprehended through biology alone. Comprehension also requires physics. When contemplating the theory of evolution, people almost universally start with an error. They assume evolution belongs exclusively to the biological sciences. Upon reading the previous sentence, some people will compound that error by assuming evolution will not be addressed in terms of science by this article – but perhaps in terms of religion. To quote from the webpage offering the “$25,000 Cosmology Prize in Evolutionary Theory” (http://cosmology.com/), “Darwin evokes or praises and makes reference to the powers of “God” and the “Creator” eleven times in his famous book and repeatedly attributes natural selection to a living “spirit” and to benevolent quasi-supernatural “powers” which keep watch over the works of the “Creator,” and which actively strives for the “good” and fights against the “bad;” and this may come as a surprise to those who never read his book, but it is not surprising given that Darwin trained to be a minister of religion.” This article consists of 3 parts – NON-DARWINIAN ORIGINS OUTLINED, CONVERTING FANTASY AND SCIENCE FICTION INTO COSMOLOGY AND SCIENCE, and SPECULATIONS CONCERNING FUTURE HUMAN EVOLUTION (the speculation is backed up with science and references).
Wotts, you are doing very well. I just made a Google search for watts up with that and you are mentioned multiple times on the first page. There I also found a page where the Dragon Slayers complained about Anthony Watts, because greenhouse gasses do not exist. The cold atmosphere cannot heat the warm surface after all. 🙂 A great evening. Everyone should laugh more.
I am also less frustrated maybe because I feel that the transition to a renewable society is making good progress. In Germany it is even going a bit too fast. Soon we might even have to close down old conventional power plants and that would be inefficient. Renewable energy is doing well and the larger the market becomes, the easier the transition becomes. The only part that worries me is that we work so little on saving energy. Doing less is less popular as doing more.
How can Stockwell simultaneously argue for low ocean heat uptake and a cumulative solar forcing driving modern warming? Where is the energy being stored?
Not that this matters given the implied climate sensitivity pointed out by richard telford above. That was, as they say, enough for me!
On the subject of solar variability and its climatological effects, I meant to link to Mike Lockwood’s corrections to recent confusion about what his research suggests, which is that a Maunder-type minimum would not significantly offset C21st warming. Since this is the obverse of the argument Stockwell advances, it is further evidence that he is mistaken.
I had a long-running argument with cohenite over the Stockwell claims some time ago over at JoNova.
“What he fails to account for is that any fixed offset in TSI, or magnetically throttled GCR’s, etc., would be balanced by a finite change in temperature – increasing radiation to space and cancelling the forcing.
Only a continuing change in forcing (total offset) can account for a continuing change in temperature, and a fixed offset in estimated TSI cannot do that. It’s a silly argument on Stockwell’s part…”
And, on that same thread, with Stockwell himself…. I find his reasoning unconvincing,
That is a really great word, Victor. The Germans have some interesting words that have no equivalent in English. Schadenfreude (pleasure derived from someone else’s misfortune) is another example. The Japanese have some unique words too. For example, the word Bakushan means a woman who is pretty from the rear but ugly from the front. Can you believe there’s a word for that?
I wish I shared your positive outlook for renewable energy. I read a fair amount of Australian news and they’ve got the highest CO2 emissions per capita in the world, they’ve just elected a PM who thinks climate science is crap and he is now trying his very best to demolish the carbon tax.
The basic point about GCM’s however is correct. There is no expectation by those familiar with this type of modeling that they are anything better than fancy conservation of energy models given the large numerical dissipation on the course spatial grids used. This is pretty well documented over at James’ where there are some references to the literature. Basically, these issues are all very well known in the numerical community and have rigorous mathematical theory behind them.
David, so what is the consequence of the large numerical dissipation? There are quite a number of highly dissipative codes that people use. That, in itself, doesn’t mean you can’t use them. You need to know if the highly dissipative nature is relevant for what you want to model. To be clear, though, I don’t know if these codes really are highly dissipative, so I’ll take your word on that for now.
KR what you say is what irritates me about this type of thing. If, for some reason, we were 1.2 degrees below equilibrium in the mid 1800s, then solar forcing could indeed be the cause for the rising temperature. The problem is, though, that if you calculate how long this should take based on the heat content of the climate system it’s around a decade (I think). So why is it taking 200 years? On the other hand, as you say, if our surface temperature is at equilibrium, then it would radiate any excess energy back into space quite quickly. So, I agree, basic physics tells you that Stockwell’s idea doesn’t really make any sense. It’s also amazing that we’re still having to explain this to people. It is relatively simple and shouldn’t be something that people are still considering as being possible.
Oh, I see. So misogynistic, sexist stuff is ok now?
Shub, where is that?
Shub is probably referring to my comment about a Japanese word. It is a bit off-topic but sexist? Really? Can’t we just all get along Shub and have an interesting discussion?
Oh, I see. Unless you were directing that word at an individual, then regarding it as misogynistic and sexist to point out that the word exists seems a little over the top. Although, the be fair to Shub, the existence of a word like Bakushan is entertaining, but actually using it would be rather unpleasant 🙂
I wonder whether there’s a German word for when your children embarrass you without feeling embarrassed themselves – kinderschämen. And then when the children become teenagers it turns into elternschämen.
And no, I have not directed the word bakushan at anybody ever.
I would say that “cringeworthy” is fairly equivalent to “fremdschämen”.
Lubos Motl, a polymath? I remember a blog post of his, where he tried to link some temperature trend in winter to the latent heat of snow. He miserably failed to correctly calculate the area of a spherical cap from 45° up. When I mentionned the error in the comments, my comment was not published, of course, and he just changed the numbers in his calculations. But he aknowledged and dismissed the error in a comment at ClimateProgress (http://thinkprogress.org/climate/2010/01/18/205363/so-its-in-the-50s-in-dc-again-and-the-global-temperature-is-still-breaking-records-for-january/ : my comment is #13, his answer #15, and a further remark by me at #36).
Raoul, I’m always a little reluctant to mock others in case I do something worth mocking (quite possible, I would imagine) but there do seem to be a number of people involved in the climate science/global warming debate who have higher impressions of their own brilliance than would seem warranted based on what they’ve said publicly.
Anivegmin, indeed “cringeworthy” would seem to be pretty much the English equivalent of “fremdschämen”.
Wotts, fair enough. But Motl is using and abusing of his MIT credentials to establish his supposed authority, so I think it’s fair to point out when he makes an error that any 14 yo wouldn’t make. And more so when that error invalidates his whole hypothesis.
Raoul, apologies if you thought I was criticising you. I wasn’t meaning that in any way. Pointing out these ridiculous errors is precisely what I’m trying to do here. I just try to avoid openly mocking those who makes these errors – although I don’t always succeed. On the other hand, I’m in no way suggesting that others shouldn’t feel free to use mockery if that works for them 🙂
> So misogynmoderating st stuff is ok now?
Shub plays the ref yet again.
Willard, yes, but I haven’t yet added a rule about not playing the ref, so technically Shub passes moderation.
Now, maybe this is obviously basic math, but when you integrate over some increasing function using the average as the baseline (as Mr. Stockwell says he does: “The parameters of the model are calculated by regressing the integral of the solar anomaly against the temperature, where the anomaly is calculated as the raw value minus the mean value of irradiance (or sunspot count) over a sufficiently long period of time.”) the resulting cumulative function is always roughly U-shaped: parabola, if the original trend is linear, or V-shaped if there is a step change, etc. He then has cut the beginning downslope off from the beginning of his graph shown above in the blog post – the whole 20th century U-shape is shown in one of his viXra papers (“Key Evidence for the Accumulative Model of High Solar Influence on Global Temperature”) and it fits very poorly before 1950 or so.
Also, in that Fig. 2 of “Key Evidence…” I’m not exactly convinced that he really uses the average as the baseline, or he has cut some of the record off from the beginning of that graph too, because the cumulative TSI goes up more than it goes down. The areas above and below TSI average should be the same and the integral should thus start and end at the same value, no? (Please correct my rusty math if necessary.) Well, obviously the temperature won’t. What the TSI baseline is exactly is not mentioned, only that the TSI data is from Lean (2001).
David says in his paper: “The fit of the model deteriorates in the early 20th century where the data is more uncertain.” Err… Yeah, that must be it.
This whole cumulative TSI business sounds very much just curve fitting exercise to me – just by choosing the start dates of TSI and temperature datasets and scaling appropriately you’ll always get a nice trend upwards in the latter part of the record which coincides nicely with the recently increasing temperatures, only because TSI is slightly higher now than it has been for some 400 years.
Tapani L., I haven’t looked at his calculation i detail, but it certainly appears like just a curve-fitting exercise. Also, as you say, it does seem as though he can always choose suitable parameters so as to fir the temperature data. There’s certainly no physics there.
Your comment is disingenuous. You merely regurgitate one of your many talking points, on each of which you’ve shown to be wrong or at least to be severely misguided. Shall we go figureing over at James’ place? Your comment not only demonstrates that you are lacking the basic understanding of the problem in question, but, more importantly, that you are not willing to learn more about it. In essence, what you keep saying is: Climate modellers are just stupid. Still wondering why some people don’t take your objections serious?
I only realize now that David Stockwell is an expert in many fields of climatology. In my beginning idealistic days on WUWT, I had a long discussion with him and a lot of WUWT regulars that were defending him on homogenization.
Like in an action movie, surrounded by dozens of enemies, the hero keeps standing. The Watties may see it as a zombie movie, you keep on killing the guy and he keeps on talking. That was fun. 🙂
Those were the times that my comments were still immediately published. Now they hang in moderation for some hours and no one notices them anymore when they finally do appear.
Oh wow, Victor. I can’t even read all of that. They’re like a pack of lions circling you and hoping to rip you to shreds. But to your credit, you are calm and cool throughout and manage to stay in one piece. Although I had to stop reading so I’m not sure how it ends.
Victor, I only managed to get about half way down the comments, but I was impressed. Remarkable self-control. Do you actually still comment on WUWT?
Rachel, I would say, it was a happy end. Or maybe an open end, it is hard to get a concrete answer, no matter how clear and detailed the question is.
I think what makes it easier to stay calm is to realize that they do not know you at all. And that it is all just ASCII, my life is not in danger. I am very happy though, that in Germany freedom of research is in the constitution and they cannot retaliate with FOI harassment.
All the stuff they through at my head, isn’t really for me, but for their enemy, reality. In a way I almost enjoy such digressions, there is no better way to demonstrate that you lost the discussion as commenting: “People are suffering and many deaths can be blamed on policies created as a result of the IPCC. Would you like proof of that? Do you feel any amount of shame for helping promote death?”
Wotts, I do not comment much there anymore. It is mainly entertainment for the lions. Writing blog posts is much more effective. It is easier to link to blog posts as to a long comment threat. And these discussions have shown me, that it is impossible to come to an agreement with a climate ostrich, no matter how small the topic is, no matter how well defined, no matter how easy it is to check, which answer is right.
I had a similar nice discussion here (where also the above insult comes from). A paper had claimed that they could not compute a trend (data too inhomogeneous), a guest post took one sentence of the article out of context that suggested that there was no trend. All you had to do is read the entire paragraph and you could see that the claim was wrong, which I wrote a post about. No one checked or at least no one admitted to have checked, no one complained about being misinformed and it was so clear. It is hopeless. I like such clear, simple (non-scientific) cases. It is a bit nit picking, but it also very clearly demonstrates to any sane person what a hopeless bunch WUWT is.
If I do comment nowadays, I get into the moderation cue. When the comment finally appears the discussion is a few dozen comments further and nobody notices, or at least they are free to act as if they do not notice. That basically kills the discussions.
Karsten, Despite your hostile tone, I am glad you read my comment, which you seem to have not really understood. Climate modelers are NOT stupid. Perhaps the problem is related to this one:
In any case, Von Neumann, Peter Lax, and Robert Richtmyer were real people and numerical analysis of partial differential equations is a real and rigorous field. I know you saw the details at James’ but perhaps didn’t have the time to actually read anything. I know your field is aerosols, so this would be heavy slogging for you and unless you have strong math, will be impenetrable.
As I mentioned there I did skim the lengthy report on the NCAR climate model and was dismayed to learn that their basic numerics have changed little since I was a graduate student. I won’t tell you how long ago that was, but it was not recent.
I still am waiting for your response to our last exchange at James’. There I asked you how denying that there are serious problems in climate science does anything to advance science. Certainly blanket rhetoric with no technical details doesn’t help me or any genuinely neutral observer.
I am actually gratified that recently there has been more discussion and papers appearing about lack of skill of climate models. In fact, it seems that as models become more complex, uncertainty is not decreasing. Some of us saw this coming.
I know you are young and inexperienced. There is actually a real scientific and mathematical debate around solving the Navier-Stokes equations on computers. Don’t get too worked up over it. I think the truth will probably emerge.
David, can you provide some kind of link for the discussion at James’s? I don’t actually know which James you’re referring to. Actually, maybe you could also explain which comment you’re responding to, as I don’t think Karsten has commented here (okay, I got caught out by the K.a.r.s.t.e.n.). Also, being condescending doesn’t, technically, violate my moderation policy, but it is discouraged.
James Annan, and DA’s recollection of the thread I believe he is talking about is very different to mine. Karsten may also be puzzled by this.
I may have just found the thread and am working my way through it. Currently, I too am puzzled.
There are actually 2 of them there. In both cases the real technical meat comes toward the end. I think one of them might be “more on that senstiivity paper.” but I don’t keep exact records of my comments on blogs for posterity or to flatter myself. There is a scientific literature for that. Do not be caught in the one discussing Nic Lewis’ and Otto et al where there is a discussion of aerosols. There’s not much there unless my memory is failing.
Wotts, I am sorry if I was condescending to Karsten. He however was quite hostile himself. And I think he knows full well that his characterization of my point is just wrong.
James Annan: More on that recent sensitivity paper.
OT, but perhaps Wotts you have some thoughts on the recent Science paper on Holocene marine sediments and the Mann response. Looks like there is a real controversy still about the climate history of the Holocene.
That Economist article, links to another Economists article, which I had already critiqued before:
I wonder whether this Economist article on the quality of science is well researched itself.
It mentions a Science study, that send out fake erroneous manuscripts to scientific journals. It claims that half of the journals accepted the paper, but without mentioning that a large part of the manuscripts were send to a list of journals known to be bad. That has quite an influence on the number.
It mentions the recent spread-sheet problem in economics. “He tried to replicate results on growth and austerity by two economists, Carmen Reinhart and Kenneth Rogoff, and found that their paper contained various errors, including one in the use of a spreadsheet.” For balance one would expect that the article mentions that the result were only quantitatively wrong, but, as far as I know, not qualitatively.
And their calculation example with the number of false positives does not convince me. Yes, scientists would like to find relationships that are “unlikely”, much less likely even as just 10%, but you find these theories not by random testing, but guided by theory. Purely empirical shotgun research is very unproductive.
When I see three problems, even if they are not that big, I wonder how many others there are. I would not let this article pass review. 🙂 If one Economist article is not reliable the other on is likely similar.
David Young, the economist article you linked to makes the claim: “Last year researchers at one biotech firm, Amgen, found they could reproduce just six of 53 “landmark” studies in cancer research. Earlier, a group at Bayer, a drug company, managed to repeat just a quarter of 67 similarly important papers.”
May I ask you, if your doctor would recommend you a certain treatment, would you trust that? If you really believe the article you linked, you would not. They suggest that in less than half of the cases your treatment would be right. Better do the opposite as what those fraudulent scientists propose, right?
Or do you realize, when you go to the doctor, that that advice is not based on a single study, but on a huge volume of interconnected studies. That that removes the influence of such bad studies. Just like the basic tenets of climate science.
Bad studies are bad for scientific progress, it slows it down. Still conclusions reached by larger communities of researchers, especially ones based on multiple lines of research are reliable.
Wotts, I wonder if you noticed William Connolley’s comment, right at the end of the thread at JA’s linked above? Right at the end.
David Young, related to your last Holocene request. Why don’t you do some research yourself, come up with a specific question and not just a vague allegation and then ask that question at a blog on that topic. Would be more productive, wouldn’t it?
First, we must be careful not to confuse Tropical IndoEuropean intermediate waters between ~500m – 100m depth with the global ocean 0 – 2000m depth.
Then we need to remind ourselves what F13’s abstract says about the variability of global land surface temperature:
Finally we need to remember that arguing that F13 indicates substantial variability in Holocene climate is to argue for a moderately sensitive climate system. Something that would preclude low values for S such as those you espouse.
You ought to be rejecting F13, not embracing it.
Victor, The Economist editorial is perhaps not the entire story, but I believe it based on what my brother tells me about the medical literature. Basically, there is a lot of poor research out there. My brother is an MD and his advice would be to do the research yourself or find an honest doctor who will give you the whole story. There are tremendous amounts of money wasted on medical treatments that have no real benefit, including many that are backed by doctors themselves, for example, vertebraeplasty. At least in medicine, they recognize they have a growing problem with bad science. There was a report in the New York Times I believe this spring about these problems. They interviewed an editor of a prestigious infectious medicine journal about his growing retraction rate. He anyway sees a growing systemic problem.
More thinly veiled insinuations that there is a “problem” with “bad” climate science. Old and tired.
I had to look up “vertebraeplasty”. That is a specialty topic, in my view. There will definitely be problems in minor fields within climatology. The community using detrended fluctuation analysis to study long range dependence (LRD) in climate data would come to mind. During a conference, someone asked a presenter using this method whether he had read Karl Popper. And his supervisor answered (weird in itself): “First of all you have to believe in LRD.”
You are right that medicine is not the best example, still even in medicine most pills do what they are supposed to do, especially if they are for a common decease and on the market since at least a decade. My asthma inhaler worked and that is no coincidence.
In Germany we have a funny situation. There is a two-class health insurance system. The poor get the old proven treatments, while the rich have their private insurances and are used as guinea pigs for latest badly validated treatments.
There is no reason to doubt the basic tenets of climate science. If you do, please have some strong arguments and not just the Economist.
Sorry, typo above:
“~500m – 100m depth” should be ~500m – 1000m.
“There I asked you how denying that there are serious problems in climate science does anything to advance science. “
Surely neither accidentally nor intentionally in the same sense as the famous question “Are you still beating your spouse?”
The vague nature of David’s question is bothersome.
Victor, I do not doubt the “basics” of climate science, for example the greenhouse effect. But the problem here is that there are a lot of issues with the way genuine controversies have been dealt with. The tropical tropospheric hotspot comes to mind. James Annan, I think also has commented on some of the other issues. The real problem I think is that the feedbacks are still not that well constrained and they depend on the dynamics of the system, which will be damped in climate models.
One thing that is good is the recent appearance of the climate dialog web site where some of these controversies are aired with disagreement experts. There was a recent one on the tropospheric hot spot. That dialogue gives a totally different picture than Santer’s paper on the subject, which came into print in a rather unconventional way it appears.
BBD – The main problem with the Rosenthal paper is that they fail to consider the effects of obliquity on the ocean circulation. Their simplistic assumption is that the circulation has remained pretty much as it is now for the duration of the Holocene.
A few older climate-modelling papers indicate that this in unlikely, and it has to do with the way that obliquity (Earth’s axial tilt relative to it’s plane of orbit) changes the wind-driven mixing of heat into the deeper layers, and with the speed of the wind-driven ocean circulation itself. Difficult to explain without graphics or animations, but models imply that the tropical thermocline warmed as obliquity increased up to the Holocene Climatic Optimum (HCO), and then cooled as obliquity decreased.
I don’t doubt for a minute that the climate cooled from the HCO onwards, or that the forams may record this general warming up to the HCO, and then decline afterwards, but much of this is related to obliquity-induced changes of the circulation, not the actual heat content of the Pacific Ocean itself. The tropical ocean surface was actually almost 1°C cooler than present-day during the HCO , despite the tropical thermocline being over 1°C warmer. Same Pacific ocean, but in a very different state compared to today.
Rob, do you mean precessional forcing rather than obliquity? Otherwise, yes.
Which brings us to the PETM. How does that fit with your thinking?
Right, climate modellers aren’t stupid. If true, why is it then that you felt compelled to make the following statements (list not exhaustive)?
“The doctrine that seems to me to be the basis of this whole thread and the previous one is the uniqueness of climate as a function of forcings.”
“Paul Williams said basically that he was focused on getting modelers to pay attention. And that’s the problem. Paul has some very good work that should be very interesting to them.”
“Gerry Browning is making some of the same points and he’s worked in climate for 30 years. He’s rigorous and trustworthy and smart .”
Read: “[…] He’s rigorous and trustworthy and smart.” … while anyone else isn’t. At least that’s how most people (who work in the field) would inevitably conceive such drivel. Out of context? Well, I don’t think so (source: julesandjames.blogspot.co.uk).
For the time being, I’m not gonna answer any of your question for 3 reasons, well, actually 4:
(1) I’m not a fluid dynamics guy (though I’m a climate/NWP modeller myself)
(2) You don’t know how climate modelling works (but aren’t shy of wild accusations)
(3) You don’t take me serious (as evident in one of your last comments)
(4) I’m not Tamsin and hence less forgiving (which I very much apologize for)
If you want me to listen again, please provide strong evidence that (1) the issue you raised is as important as you claim it to be and (2) that it has been deliberately ignored by those in charge.
Ok, Karsten. Your content free response tells me only one thing: that you don’t like me very much, but I already knew that. 🙂 Since you don’t know me and can’t comment on the substance of what I said, that seems rather odd to me. Maybe another symptom of a politicized field?
“…which came into print in a rather unconventional way it appears.
Dripping, again. Needlessly distracting.
Where did I say I didn’t like you? I don’t like your through and through distrustful tone. That’s all!
Speaking of substance … short on evidence?
Karsten, Maybe I judged you too harshly. My distrustful tone is based on the growing evidence that science has some basic problems. There was an editorial in Nature on this, the Economist editorial and much longer paper, the article in the New York Times, and my brother who has known about these problems for at least 15 years. Vioxx and vertebraeplasty are not isolated instances. The medical literature is shot through with anecdotal studies, conflicts of interests, and the influence of money. And that’s not just the drug companies either.
Science is about trusting but verifying. Recently, there has been too much trusting and not enough verifying. That is becoming more and more obvious except to people who have a political interest in not admitting there is a problem.
In the climate science arena, the politization is probably what is behind people not wanting to admit any real problems. They are afraid of what the “enemy” will do with these things. That happened to James a couple of times, mostly from non scientists though. But Jules took some strong and ridiculous lambasting for a single phrase in an excellent interview. By the way, the offending phrase was in fact absolutely true and unremarkable. Then Von Storch showed up and corrected the record. When people are too fanatical about denying any problems, its not a good sign.
I can only hope for the chance to apply that bad Japanese word to DY. Very soon now, with any luck. Our host may find this tool handy (h/t Eli).
“Rob, do you mean precessional forcing rather than obliquity? Otherwise, yes.”
BBD – I’m not suggesting that precession (the long-term wobble in Earth’s rotation [I just add this stuff for the casual reader’s benefit]) is unimportant for Earth’s climate throughout the Holocene. Very clearly it was, for instance by shifting perihelion (Earth’s closest point to the sun) from the Northern Hemisphere summer (at the Holocene Climatic Optimum – HCO) to coincide with the Southern Hemisphere summer (today). But that’s not the key driver for the ocean circulation, obliquity is. Obliquity (the central axis of this long-term wobble [precession]) decreased 1° over the last 9000 years, meaning that the tropical oceans are now closer to the sun throughout the year.
This has consequences for the subtropical ocean gyres because the greater warming of the tropics means greater uplift at the equator and therefore stronger easterly trade winds which (partly) power the ocean gyres (the mid-latitude westerlies strengthen too). Move toward greater obliquity and the gyre circulation weakens and thus the transport of heat in the thermocline, out of the tropics, weakens. The subsurface is warmer than present (at the HCO) despite the cooler sea surface then. The sea surface at higher latitudes is warmer because of the greater seasonality. Move forward in time from the HCO and the export of heat out of the tropics begins to strengthen as the ocean gyres spin up. The higher latitude sea surface (where the water flowing through the Indonesian ocean pathways partly originate from) cools.
Long story short, those forams are not only recording the long-term cooling of the ocean since the HCO, but also a local effect due to changes in circulation. As it stands, they are unlikely to be a reliable recorder of the Pacific Ocean heat content.
David Young, what I find quite intriguing is that you refer to Santer et al’s supposedly “unconventional way” of getting published, but fail to notice the rather “unconventional way” that Douglass et al analyzed the data (to which Santer et al was a response). The “unconventional way” that Santer et al was published had everything to do with the basic flaws in the Douglass et al study.
Since you appear to be so concerned about the various problems in scientific reporting, I would have expected you to laud the Editor for correcting the literature and not letting the flaws in Douglass et al muddy the waters.
David Young, if you’d read my moderation page you should have noticed that I’m discouraging discussion related to the behaviour/trusthworthiness of climate scientists (unless you can provide some fairly convincing evidence). Can we go back to your very first comment. You mentioned dissipation at the grid scale. This presents two possible problem. If you set your artificial viscosity terms badly, then your code will try to resolve shocks on a scale smaller than the grid scale and will likely not conserve energy. The other, as far as I understand it, is that all grid codes will dissipate structures at the grid scale. This is only a problem if you expect to have lots of power at small scales. Can you explain why – if it is – an issue for GCMs.
Rob, that’s very interesting. Thanks.
David Young, your comments seem to be uniformly vague and often entirely rather unrelated to the issues (e.g. your vague insinuations about “problems in science” in relation to biomedical research).
Why not try to be scientific and specific. e.g. you have said:
1. ….”the recent Science paper on Holocene marine sediments and the Mann response. Looks like there is a real controversy still about the climate history of the Holocene.”
Fine. So give us your thoughts on this paper. What do you think it says, what does it mean and what is its significance wrt our understanding of Holocene temperatures? Be specific.
2. “That dialogue gives a totally different picture than Santer’s paper on the subject, which came into print in a rather unconventional way it appears.”
Good. Give us your thoughts on this dialogue. What “picture” do you consider it gives and how this differs from that derived from “Santer’s paper”.
Thanks Rob – that’s opened this up for me considerably. Much appreciated.
Why have you ignored my earlier question?
You have plenty to say here, so talk to me.
Or, while he’s at it, the mid-Piacenzian Warm Period. Hard to reconcile with low sensitivity. He *should* be interested since the models fail at both of them.
Well, Steve gets it. So what about DY?
How do we get a GHG-triggered hyperthermal if we’ve over-estimated the feedbacks to GHG forcing and models over-sensitive as a consequence?
Rob Painting re your comments above: https://wottsupwiththatblog.wordpress.com/2013/11/01/watt-about-david-stockwell/#comment-7914
I don’t see that there is any particular “problem” with Rosenthal et al [Science 342, 617 (2013); published last Friday]. The paper is a report of reconstructed ocean heat content (OHC) based on Mg/Ca ratio based temperature proxy in foraminifera in several sediment cores from two areas in the Western Pacific. I imagine there could be many problems with some of the interpretations of the paper since I expect it will appear as catnip to climate science misrepresenters! But the paper itself doesn’t seen very uncontroversial.
For example the authors are quite clear about the limitations of their analysis wrt global OHC [” With no additional IWT records, it is difficult to assess the global extent of the trends we have reconstructed”; “…the ocean heat changes apply only to the Western Pacific ( ~ 25% Pacific volume”)]…they point out that the ocean heat content (OHC) in their sample areas is increasing very rapidly now [ “The modern rate of Pacific OHC change is, however, the highest in the past 10,000 years..”]…the describe the fact that their proxies indicate no significant increase in Intermediate Water Temperatures (IWT) between 1850 and 1950, despite the fact that direct measures of Pacific sea surface temperatures demonstrate an uncontroversial rise in temperature during this period (indicating that either their OHC proxies are flawed, or more likely that the local IWT-based OHC is not representative of the Pacic as a whole or that the IWT OHC is uncoupled from the sea surface temperature – and the authors discuss the latter).
And so on. It seems a perfectly fine paper. Have to be careful not to consider that every paper needs to be interpreted as supporting or contradicting tedious set piece “arguments” in the blogosphere!
Yes, Wotts, one problem here is that any subgrid scales will of course be invisible unless there is a subgrid model. There is an interesting theory about this separation of scales issue. It’s due to Prof. Thomas J. R. Hughes of Univ. of Texas Austin and is a generalization of his SUPG method. Unfortunately, to properly account for the subgrid scales, the Green’s function for the operator is required and that is only known for simple linear operators generally.
The fundamental theorem of numerical analysis is that a numerical scheme will converge to the “correct” continuous answer if its stable and “consistent” which basically means that the truncation error vanishes as grid size goes to zero. If its unstable, you have spurious oscillations. Stability requires numerical dissipation at least for most interesting models of fluid flows. Generally, this dissipation is second order in the local grid spacing. Generally, the goal is to minimize this dissipation while retaining stability. Now the consequence of this is that if the grid is too course, dissipation is far too large, damping the dynamics, for example vortices will be smeared and dissipated far too quickly, layers likewise will be dramatically smeared, shock waves (which are not an issue in the atmosphere) will just disappear, etc. But more importantly, the overall forces, even on the largest scales will be wrong.
The problem here is that climate models are just weather models run on a much coarser spatial grid and with a big time step. So one would expect a lot of the dynamics to disappear unless accounted for by subgrid models, which are the source of huge uncertainty. The problem of subgrid models is complex. These are usually just themselves added viscosity that is a complex function of other characteristics of the flow. Everyone acknowledges that this is a potentially large source of errors. And there was just a summary paper for the NASA aeroelastic prediction workshop that says it very nicely. It’s not what a lot of people were expecting. Basically, the time averaging of common subgrid models does not adequately reflect reality in separated flows.
Now Gerry Browing makes the point that in fact, these usual sources of unphysical dissipation aren’t the only ones. There are explicitly added dissipation terms to stabilize the method. One source of this is the filter commonly used with the leapfrog scheme as Paul Williams pointed out. There is also something called hyper viscosity.
It’s a complex issue, suffice it to say that understanding of Navier-Stokes simulations is in its infancy. Recent work is actually negative in that it indicates our earlier naive beliefs are not supportable. And its about time. Just as in climate, in fluid dynamics, there has been at work a triumphalist technocratic unspoken assumption: If we just add all the physics and get a big enough computer, everything will be good. Like a lot of triumphalist doctrine, this one is now meeting reality.
I’m not sure what your moderation policy has to do with what I’ve said here about climate science. I know its a dogmatic issue among the “anti denier” camp that there are no fundamental problems in climate science and there are infinitely complex blog posts devoted to climate science apologetics. I have not accussed anyone of dishonesty. I can tell from the clientele here that there will be a lot of noise and sniping because most of your commenters are fully committed to the cause (as they see it). I’m just hoping for an interesting dialogue with you and the few others who have something of substance to say. On numerical models, that’s a very small proportion of the climate science community (who I get the feeling are mostly just users of the codes) and on even the best blogs, its virtually zero. The codes themselves were in most cases written a long time ago for weather modeling. I do learn something from your posts and comments though.
Chris – you don’t appear to have understood the discussion above. The authors can’t have it both ways, the forams are either reliable recorders of Pacific Ocean heat content, or they’re not. This is kind of crucial to the whole interpretation of their paper.
That the authors don’t appear to understand the ocean circulatory changes that (likely) took place from the HCO onwards, due to orbital/rotational factors, is not encouraging. I note in the interview with Andy Revkin, that Rosethal talks about the climate modelling/oceanographic community will need to look at their data. It struck me as rather strange that the authors don’t appear to have consulted the already existing scientific literature on modelling of ocean change throughout the Holocene – there’s plenty of it.
David, I guess Wotts’ reference to the moderation policy has to do with your poorly hidden claims of malfeasance. You come with one right after claiming you have not accused anyone of dishonesty, by stating that most of the commenters here are “committed to the cause”. Your reference to Santer’s paper as having had an “unconventional way” of getting published is also telling. I think it is time for you to consider your own bias.
David, as far as moderation goes, it’s simply that I don’t really see how comments that imply that there are issues with the scientific credibility of climate scientists are constructive. I’ve read Ben Goldacres’s book, for example, which discusses problem with the pharmaceutical industry, but that doesn’t immediately imply a problem with climate science. No science area is going to be perfect, but I find discussion about scientific honesty simply a distraction and not constructive. We’re certainly not going to achieve anything if we do discuss it.
I’m slightly confused about your view of hydrodynamical simulations. It’s not really in its infancy. It’s quite a highly developed field. I’m also slightly confused by what you mean by convergence in this context. If I consider a time-independent calculation, then I can run it till I reach some kind of converged solution. GCMs are, as far as I’m aware, time-dependent calculations and so they’re not being run to a converged solution. Convergence is established by considering how the result depends on resolution (although there may be sub-grid calculations that settle to a converged state). If you get a different solution for different resolutions, then that would indicate that you simulation hasn’t converged. I don’t know if climate scientists have done convergence studies, but I’d be surprised if it hadn’t been considered.
As far as dissipation goes, sure all codes will have problems with artificial dissipation which should be minimised. Again, I’d be surprised if this hadn’t been considered by the climate modellers. The other issue though, is that the fundamentals of AGW don’t require GCMs. GCMs have various functions. One would be to try and understand how the climate will change as anthropogenic forcings increase. Another would be to provide estimates for the ECS and TCR. However, much of the fundamentals of AGW are basic physics. Radiative transfer, for example. If you’re basing your views of climate models on the “hiatus” then I think you’re over-interpreting this mismatch. Firstly, we’re interested in long-term trends. Secondly, multi-model means are likely to smooth out these short-term variations and so it’s not that surprising that there’s been a reasonably short period where the models and observations are not a perfect match.
Hi Rob…yes I understand the discussion above thanks. You’re suggesting, as the authors also allude, to the possibility/likelihood that the measurements of OHC in the areas of the Western Pacific sampled don’t represent the variations in OHC across the entire Pacific (let alone globally) due to changes in the disposition of the oceans driven by ocean circulation changes arising from Earth orbital parameter “cycling” throughout the Holocene.
That’s certainly a possibility. For now I am quite relaxed about taking their data at face value – Variations in proxy temperature extrapolated to changes in ocean heat content in a manner that the authors are quite clear about (one can choose to accept their extrapolations or not). The authors don’t come to any fundamental conclusions about their data in relation to climatic progression/global ocean heat storage/response to forcings etc. that is of direct relevance to global warming, other than to point out (i) the useful idea of the oceans as heat capacitors (we knew that already), (ii) the divergence of apparent changes in Pacific OHC with changes in SST during 1850-1950 and (iii) the interpretation that OHC in the Pacific is rising faster than at any time in the last 10000 years.
I think it’s a value study, and likely to stimulate more extensive efforts to extend this data set. It’s obviously of great interest to modellers and the oceanographic community, since it provides potentially useful local experimental information as targets for extended understanding of the ocean system.
“…climate models are just weather models run on a much coarser spatial grid and with a big time step.”
Hyperbole isn’t helpful. To a layperson David’s explanation sounds informative but when a layperson armed with a little bit of knowledge suddenly encounters a lump of obvious rhetoric embedded in what otherwise seems reliable and possibly helpful to better understanding, the reliability index of the whole effort slumps.
For my part, when I ran into the weather model snark, I stopped listening. Maybe I’ll miss something important but I think it’s more efficient to stick with teachers who are not freighted with extraneous baggage. The weather remark is a baggage tag.
There’s so much emotional energy invested in this affair that it’s very difficult to return to a purely substantive discussion. David’s not alone in that, of course.
Re: David Young, I always find it informative when someone refuses to answer a question.
Further to my last remark, I can only think of a single website that discusses climate science fundamentals without ever swerving into irrelevancies, that being Science of Doom. Naturally the author isn’t able to maintain complete control of comments without completely shutting down discussion, but the articles themselves are entirely devoid of rhetorical flourishes and all the rest.
Of course, SOD is deemed “too boring,” a revealing assessment.
BTW, I’m not trying to offer negative criticism of Wott’s effort here. WoUWT seems to me a refreshing attempt to have some stimulating and productive discussion while not letting things slide completely into the ditch.
Do you know who writes the Science of Doom site? I’ve read a bit of it and it does seem very good but I can see that “boring” is probably a fair assessment.
Convergence with respect to time step size has been studied by Paul Williams. It can have a dramatic effect not just on short term trajectories by on the long term “climate” as well. He has a great video at the Newton Institute. There is something about this in the NASA aeroelastic workshop summary paper. Solutions ARE sensitive to time step size selection for separated flows, but most participants in the workshop DID NOT study this issue, presumably because its too computationally expensive. We can’t really achieve blind unique steady state grid convergence for a simple attached flow as the AIAA drag prediction workshops showed. But your point is a good one. How do you measure accuracy of a time dependent calculation. Typically, you measure accuracy of the answer at a fixed time point. You could use the H1 norm for the entire time trajectory. In these standard mathematical measures, most time accurate calculations are not very good. So, some want to define something like the “climate of the attractor” and that may be more intesting. The problem is its dreadfully difficult to prove anything or do rigorous computational experiments with this vague concept. This issue is really a theoretically open issue and very little is known.
What I sense here, and its OK, is that you see the field from the outside and everything looks under control and well quantified. My point is that this is the result of some pretty heavy positive results bias at least in fluid dynamics and that the bad news is just now starting to come out as people start to apply scientific concepts like blind testing. See some of the references at James or the aeroelastic workshop summary paper. There is also a false impression given by some upbeat marketing that models are the basis for everything in modern life. This is a half truth. Models are used for example to design airplanes, but the models are simple models constrained by a century of test data. Structural sizing design is really done not with finite element models but using test data arrived at by destructive testing of many small test articles.
I do think far too much money is invested in GCM’s when perhaps better data, more fundamental understanding, and better constrained simple models have more leverage. GCM’s seem to be sufficiently complex that no one fully understands every thing about them, but many seem sure that to question them is to question the intelligence or integrity of climate science. I think recent progress in fluid dynamics is important and is finally correcting a long period of overselling and false hopes. These results are not irrelevant to climate science.
Thanks, Doug. That’s certainly the intent. Whether or not I’m achieving it or can continue to do so is another matter 🙂
David, I’m finding it hard to know how to respond to your comment. I’m partly with Doug here in that you say a lot of things that sound right but the basic message seems to be “it’s really complicated and we can’t do it properly yet”. That’s, in a sense, always true. If we applied that attitude to science we might never do anything. You still haven’t really said anything to convince me of a specific problem with GCMs. All you’ve said has been general observations without any specifics.
You say you see the field from the outside and everything looks under control and well quantified. As far as I’m aware, you’re seeing it from the outside too. I’m not suggesting there are no issues with aspects of GCMs and climate science, simply that the issues are likely no different to other complex science areas.
Is too much invested in GCMs? From the “skeptic” side that may be true. Most of the issues that they highlight seem to relate to the “hiatus” and that it wasn’t predicted. I don’t think that climate science itself relies heavily on GCMs. It is one aspect that gives some indication of what might happen in the future, but is not the be-all and end-all of climate science.
David Young, I don’t think you’re doing a terribly good job of presenting you criticism. You’re mixing together generalized complaints against biomedical science from editorial in newspapers with various diatribes against “causes” and the use of computational models in modern life, with very detailed discussion of damped dynamics due to over large numerical dissipation in over large grids within numerical climate models. I expect the latter is something you have some knowledge of from your own working career.
But you need to be more explicit in identifying what you consider to be the significant consequences with respect to climate models. Damped dynamics is not necessarily a major problem if those dynamics don’t have a large effect on the essential energy balances that define the accumulation of energy (and it’s distribution) in the climate system under the influence of a forcing. Clearly climate models have been very successful in anticipating the evolution of some fundamental elements of the climate system under the influence of anthropogenic forcing (we could discuss these). In any case (as Wotts has indicated) our understanding of the climate system, it’s response to enhanced forcing and the likely consequences don’t arise form climate models..they arise from physical understanding of the climate system. The models allow us to make future projections of likely outcomes under various future scenarios relating to enhanced emissions and forcings etc.
So can you formulate what you consider specific consequences of your perceived flaws in climate models to be?
Rachel, Science of Doom is run by Steve Carson.
Yeah, SOD under-performs on “stimulation.” If I’m honest with myself, some of my attraction to the climate change “debate” lies in overwrought histrionics. 🙂
David Young, there is no need to proof that climate models are not perfect. No one would claim so.
As you have stated, they do not conserve energy. Atmospheric models have terrain-following coordinates (otherwise, one would need a very high vertical resolution up to the height of the highest mountain to be able to model exchange processes (friction, water and heat fluxes) between the surface and the atmosphere, which is numerically too expensive). These coordinates make the equations of motion much more complicated and many small terms in the equations are neglected.
As you have stated, they are dispersive. They also have a finite top. They have a finite resolution and many atmospheric processes occur at a smaller spatial scale and are modeled in a simplified way (parameterised). Important examples are clouds and precipitation. You cannot model every cloud droplet and how millions of them grow to a rain drop in an atmospheric model. The chemical composition of the aerosols on which the cloud droplets start to grow is neglected in almost all models and always simplified. Most models have a fixed seasonal for the vegetation, in some they respond in a simplified way to the weather. And none of the models have a tractor driving trough the field to remove the wheat on a specific day. And none of the models try to model how the harvest day would change as a result of non-existing predictions for changes in EU-subsidies to farmers. Some weather prediction models simulate the influence of mountains on the insolation in the valleys, as far as I know, this is missing in global climate models, where the sun is not shaded. And I could continue for a lot longer. Climate models are not perfect and will never be perfect. The world is too complicated.
The question is not whether the models are perfect, but whether they are fit to a specific purpose. I you think the dynamics of the models are a problem for climate projections, if that is your passion, try to improve it and show that if makes a difference for climate projections. I can imagine that you would be able to improve weather predictions, especially if you have better ideas as the hundreds of mathematicians and atmospheric scientists working on the numerics, I would personally be very surprised if it made a difference for the climate sensitivity.
If it did make a difference, the atmospheric models would have the problem that their estimates no longer fit to the empirical estimates.
My personal advice would be to have a look at the parameterisations for clouds, precipitation and exchange processes. There you will much more likely find approximations that influence climate projections. I am not the only one who thinks so. So you would face more competition, but that is where you may make a difference, if you are smart enough.
Victor’s points remind me of something the head of ESRI said, when asked about a particular omission of GIS models that would be extremely difficult and not very rewarding to repair. “We already have a database for that. It’s called ‘Earth'”.
That said, I can understand how it’s bothersome that models leak. Depending on what domain a person works in, it could become a horribly uncomfortable itch.
Yes, Wotts, its complicated. I would assert the following points that are less weedy and technical.
1. Recent results in fluid dyamics indicate that the problem with simulations may be a lot worse than we thought. I gave you some references.
2. It is problematic I think that GCM’s play such a prominent role in this field given the growing documentation of their limitations. There was a paper in Nature I think pointing out that decades of “progress” and huge computational costs have not reduced uncertainty. (See point 5 below).
3. Adding “more physics” or running with finer and finer grid and smaller and smaller time steps is the naive approach to any problem with model skill and its a little like burning witches at the stake. Adding more physics can actually increase uncertainty by multiplying parameters. This is true for Reynolds’ stress turbulence models for example.
4. Climate models by and large use outdated numerical methods. GISS may be something of an exception to this rule but even they don’t use the latest advances in adaptive finite element technology. I gave you a reference for this newer technology. In the case of the leapfrog scheme, the problem is obvious and can be seen with an hours time with pencil and paper. I gave you a reference to Paul Williams’ work. I don’t know how much difference it will make, but it makes a big difference in fluid dynamics.
5. Given the recent progress toward recognizing how little we can simulate reliably, working on better mathematical theory might be helpful. There is some work on dimensions of attractors by Temam et al, but its disconnected from actual computations. We need an emergency effort to determine an appropriate norm to measure time accurate calculations and also some quantification of this doctrine of the “attractor” which Victor perhaps subscribes to, viz., that numerical errors don’t matter because in the end you get sucked into the omnipotent attractor. This has no scientific basis that I can ascertain. Paul Williams showed how this incantation is false incidently.
6. There is a problem in science generally that has been getting worse. It relates to positive results bias, the career rat race in science, the intrusion of money (and soft money is a large part of this), the lack of interest in validating other people’s results, etc. It’s symptoms are also present, I believe, in climate science, even though I expect people here to deny that in order to keep the “deniers” from having any ammunition. This view is held by a significant number of mainstream scientists, yes even climate scientists and is an issue that deserves attention.
A few things. With regards to 1), I’m having trouble finding those references. Are you referring to Paul Williams here? Which of his papers specifically claims problems with simulations?
Your second point is one you keep repeating. Wotts has said above that he does not think climate science depends heavily on GCMs. It’s no use saying “yes, they do” while other people say “no, they don’t” and around and around we go. If you want to press this point then you need to add something new to support it.
In points 5) and 6) I think you are attacking the accuracy of climate models. Correct me if I’m wrong here. You say they use outdated numerical methods and then imply that they are unreliable. I don’t really know very much about this topic so if you could provide some actual links in support of these claims I would much appreciate it. It’s my understanding that the purpose of climate models is to forecast future global temperature changes and that by and large they have been remarkably spot-on at doing this. See Global warming predictions prove accurate in the Guardian which references a recent paper by Myles Allen.
This is not to say that climate models are perfect but they are modelling something very complex and as Victor has pointed out, how clouds will behave is particularly challenging.
I’ve just had another sift through Paul William’s publications to see whether I can find the paper you refer to about problems with simulations. Is it this one: Mathematics applied to the climate system? If so, then I don’t see anything wrong with assessing climate predictions, identifying uncertainties and problems and then looking to improve their accuracy. This paper does seem to suggest to me that all is not wrong in climate science which is what you allude to in point 6). If there was overwhelming confirmation bias then we wouldn’t see papers like this which specifically look for shortfalls in our current understanding and which then try to suggest improvements.
David, Rachel’s largely responded how I would have done. Firstly, I don’t think GCMs are nearly as prominent as you seem to think. Without GCMs we’d still have extensive evidence for AGW and estimates for the ECS and TCR would be largely unchanged. Most of the rest of what you’re saying may be true but seems more like general comments about computational science than about climate science in particular. We can always improve things, so what is the particular significance for GCMs. See Victor’s point above, for example.
Your point 6, in particular, is one that I find a little frustrating in that there is an element of truth but doesn’t indicate that there’s anything special about climate science. I’m well aware that there are aspects of an academic career that are not optimal, but that doesn’t mean I would dismiss major areas of science. It’s a general issue, not one that applies to climate science in particular.
It’s sad but true that the climate science debate has more entertainment value than it really deserves. But entertainment value aside, Wotts does something here that not many others can replicate and that is that he explains things simply. I find his posts easy to read. It’s very hard to write about something complex in a simple manner.
Perhaps I have no patience or my powers of concentration are lacking because I do find my eyes glazing over sometimes when I read some of the climate science stuff on the web. I once heard a joke and I can’t remember who said it but it was along the lines of the person not having enough concentration to even read an entire tweet. Sometimes I feel a bit that way. 🙂
Rachel, I’ve spent quite a bit of time at SOD and I find my eyes glazing over, crossing, closing. 🙂 But the same thing would happen if a semester’s course in chemistry were foreshortened into 2 weeks of 10-hour lectures. There’s so -much- there; one has to know when to come up for air.
That said, what’s interesting and rather wonderful about SOD’s approach is how it illustrates how much we take on faith, when we talk about science. SOD happens to be about climate science, but leaving that specificity aside his approach seems to be to leave nothing only assumed or left buried and invisible as implicit. Employing that tactic shows how rich in detail “our” understanding is, and how easy it is to imagine we understand a topic when in reality we’re scribbling in a kind of shorthand that is only descriptive in the roughest sense.
SOD’s expositions are a humbling thing in more than one way. They’re a reminder to me of how easy it is to shoot off my mouth without having a true inkling of what I’m talking about, and as well they’re humbling because they reveal the extraordinary power of the communal effort of science.
Doug, yes you’re quite right about SOD. I’ve also struggled with some of the posts there, partly because it is indeed complicated and partly because it goes into all the gory details and requires a high degree of concentration. I do like that I can go there to check that my very basic calculations actually make sense.
Yes, there is a lot of information at SOD. When I was at University I had a lecturer who observed that students seemed to be most alert during the first 20 minutes or so of the lecture and then started to lost concentration after that. He decided that the best way to combat this was to stop the lecture halfway through and show us photographs of his holidays. It worked really well. This little interlude only lasted for a few minutes but it woke everyone up and when the teaching resumed it was as though he got that first 20 minutes back again. Maybe Steve Carson needs to include some holiday photos in his posts at SOD. 🙂
David Young: “Adding “more physics” or running with finer and finer grid and smaller and smaller time steps is the naive approach to any problem with model skill and its a little like burning witches at the stake. Adding more physics can actually increase uncertainty by multiplying parameters.”
So you want to make the model dynamics perfect, but do not care about the accuracy of the model physics? I would argue that for an accurate result, you need both. Given that computational and human resources are limited you have to balance how much work you put in the dynamics and how much in the physics.
A high resolution is just as important as dynamics and physics. It is needed to be able to model the influence of the land sea mask, of the orography (mountains) and the vegetation. All these fields display variability on all scales, the better you resolve this variability, the more accurate the result. Again, given limited computational resources, you will have to balance these advantages against using better dynamics and model physics.
Do you know whether the unspecified “modern” methods you are dreaming of, work in a parallel computing environment? Much of the recent work in mathematics on numerical methods is not suited for parallel computers. The climate system is a high dimensional system, with many degrees of freedom. Without parallel computers you can only make short or coarse model runs.
“1. Recent results in fluid dyamics indicate that the problem with simulations may be a lot worse than we thought. I gave you some references.”
You didn’t. All you did is talking about stuff Paul Williams and a few others did. It’s funny, I had the pleasure to listen to presentations given by Paul twice. I can’t remember him saying anything which came close to what you’re claiming here, not even remotely. So again, please would you be so kind to provide some evidence for your wild assertions re dissipation impact on GCM’s! Referencing to a paper which argues for potential problems isn’t gonna be enough.
Problem: You don’t have one! Another problem: As wotts already said, you’re an outsider to the field (which is easy to spot for anyone who works in the field) and should therefore be much more careful with your choice of words. Telling the experts what is known or not known in their field is nothing more than audacious arrogance. Telling the experts what they need to do or not won’t make them listen, even if you had something interesting to say (which you don’t).
Victor, Yes finite element methods are indeed very parallel. Adaptivity and error estimation are likewise quite parallel. The paper I referenced at James discusses this issue. With proper programming they can get speedups of 5000 on 10000 cores or even more. This is documented in 1000’s of papers in computational science and engineering over the last 15 years. Look for Thomas Hughes, SUPG. I don’t know if this technology will help, but it can’t be worse than leapfrog and non-conservative finite volume. Discrete conservation is guaranteed for these methods, so there are no problems with conservation of energy.
I agree that many believe that complex (higher fidelity) physics + fine grid + small time step = right answer. That is simply untrue. One reason is that the more physics (subgrid models) you add the more parameters you introduce and that increases uncertainty if you don’t have enough good data to constrain them all. Often simpler models that are well constrained by data are more accurate. We have a paper in press on this. If you can somehow contact me, I can send you a copy since its been formally accepted. The problem here is that as computational complexity increases, rational error control becomes more difficult, not to mention the butterfly effect, etc. etc. (You know the sad tale) You can also look at Aeronautical Journal July 2002, pp.349 ff. This is about turbulent flow and the atmosphere is very turbulent. The editor introduces the paper this way:
“From time to time we invite acknowledged experts in a particular field to write a survey article for The Aeronautical Journal. Prof. Leschziner is such an expert in turbulent flow. …”
From the conclusions section of the paper:
“This review has provided evidence that eddy-viscosity models are fundamentally flawed and often perform poorly in flows featuring separation, strong shock-boundary-layer interaction and 3D vertical structures. More seriously, perhaps the models do not display a consistent behavior across a wide range of conditions. In relatively simple flows, which develop slowly and in which a single shear stress (expressed in wall-oriented coordinates) is wholly dominant, eddy-viscosity models can be crafted to give the correct level of this stress and thus yield adequate solutions. … the fact that many flows are an amalgam of shear layers and regions in which turbulence is dynamically uninfluential explains, in part, the moderate level of success of eddy-viscosity models.”
For the atmosphere its much worse. Not only is it turbulent, but vorticity dynamics are very important at all scales. These effects must be accounted for in some way. Adding viscosity is the time honored way. I was able to verify from the NCAR model document that they use eddy viscosity at least for the boundary layer. But they appear to have so many “viscosities” its hard to disentangle them all. There is also IFASD-2013-1A which is the workshop summary I mentioned earlier. They confirm pretty much these conclusions. Let’s not discuss LES or DES, these are decades in the future and have a lot of other problems.
My point 5 above says that what is really needed is some fundamental theoretical work on attractors, their dimension, and some careful computational work on simple sub-problems where there is at least some hope of detecting bifurcations, stability issues, etc. We need to understand how to measure the accuracy of time accurate solutions and how to quantify it as much as possible. The answer may be its not really possible, but we should try very hard. Temam el al’s recent book has some theoretical ideas, but so far they are not practical.
Rachel, look for Paul Williams’ video at the Newton Institute from perhaps 3 or 4 years ago. He goes through the leapfrog problem and shows a simple Lorentz system calculation to show that it does make a difference for the climate.
Karsten, I’m ignoring your essentially content free response which is really just an appeal to authority. You won’t understand the technical details probably, but there is a rigorous and deep theory here that makes a difference even in climate science where the experts no doubt know better. 🙂
Wotts, Its pretty obvious what the problem is with climate science. Its the same problem as in medical research. You have lots of money, strong politization of the science and the over reaction of some scientists to it. There are also I would argue problems with statistical methods, at least there are a lot of statisticians who think that’s the case. In medicine statisticians are on every important team to make sure things are done correctly. So far as I can tell, and you can correct me if I’m wrong, many prominent climate scientists have resisted this and have tried to demonize legitimate debate. Another problem I think is that blogs don’t do enough to control flaming and nasty comments by the political fringe types. There are so many overtones of cultural prejudice here too. Academia being mostly a liberal place, conservatives are particular targets of this prejudice. I haven’t seen that here though 🙂 You saw Carl C at James, a prime example of a very nasty and bigoted human being.
Rachel, I saw Allen’s paper discussed in the Guardian. I am relying on Ed Hawkins I believe is using a more reasonable baseline, one that I believe Richard Betts endorses. If Allen is right, why did the IPCC lower their likely range by 10% from the model forcasts for the next 10 years I think and assert that some GCM’s are too sensitive to greenhouse forcing? I admit this is a complex story, but the truth I believe is closer to the IPCC and Hawkins.
“So far as I can tell, and you can correct me if I’m wrong, many prominent climate scientists have resisted [statistical advice] and have tried to demonize legitimate debate.”
Correction is difficult as David has failed to supply any names of actual scientists said to have refused assistance by statisticians. Too vague to be discussed, same problem as David’s putative political pollution of climate science.
Also, medical research is famous for statistical blunders, the same blunders, over and over again. Apparently the eagle-eyed statisticians assigned to medical researchers are frequently on coffee breaks. Google “common statistical blunders in medical research.”
This conversation is so much better when it a) stays in the right domain and b) doesn’t trail off into vague innuendo.
I’ll have to retract my claim that I haven’t seen cultural prejudice here. 🙂
Medicine involves statisticians and they still get it wrong. Climate science is so much better and statisticians are unanimous on that.
Pingback: No, David Stockwell, it’s not the Sun | Wotts Up With That Blog
“In medicine statisticians are on every important team to make sure things are done correctly.”
“Medicine involves statisticians and they still get it wrong.”
Same author. What a mess.
David, screwing down my tendency to go elliptical I honestly say it is very disappointing when you begin dripping innuendo and making unanswerable accusations, because any confidence I have as a layperson in your capacity to inform me is horribly diminished when you do that. Earlier in this conversation you’ve provided some very interesting information that seems grounded in fact, but then– bang– again the confidence shattering drip of attitude evidences itself once again. If you allow your negative feelings to color your portrayal of one facet of reality, how am I to know that those feelings are not interfering with everything you write?
This blog seems to be a fresh attempt to stick with substantive discussion. It’s an opportunity for some of us to learn, if it’s allowed to reach that potential. When we’re invited or invite ourselves into a person’s home where it’s clear that napkins are expected to be placed in laps before picking up the knife and fork, it’s just plain rude not to follow those conventions. I can’t claim perfection on that account myself, but it’s nonetheless incumbent on us to try and calibrate as best we can against the expectations of our host.
I would say in general that overall quality of the medical literature has been improved as a result f professional statisticians. Do you agree? My brother who knows says this is true.
Better for everybody if you stick with what you know, David.
That doesn’t answer the question, Doug. Apologetics is not an appropriate response to a scientific question whose answer is pretty obvious even to non specialists. My brother is an MD and was director of medicine at an HMO. He is trustworthy and I actually know him. You I only know from your apologetic exploits.
David Young says:
As a reader of this exchange, I find this completely unsatisfactory. It is evasive. Furthermore, you cannot yourself accuse a commenter of “content-free” responses when you are persistently failing to provide content in your own. Despite repeated requests. It is clear evidence of bad faith.
David Young, it may well be that professional medicine has benefited from statisticians, but that doesn’t immediately imply the same would be true for climate science. I do think there is a big difference between a field where well-defined surveys may dominate and one that is based on the fundamentals of a physical science.
To go back a bit though
One reason they lowered their range is, I believe, because of the more recent observationally constrained estimates. These predict an ECS of between 2 and 2.5 degrees per doubling. However, they have certain issues (such as aerosols and internal variability) that might indicate that they’re more likely to be on the low side. Essentially, there is not much evidence that we’d expect the ECS to be much lower than earlier estimates. So, overall, the evidence still supports an ECS closer to 3 degrees than to 2. Maybe the estimate for the next 10 years should be lowered but that’s more likely because of internal variability than because we’ve overestimated the ECS.
On the subject of SoD, I’ve learned quite a lot there although admittedly some of it is a bit too advanced for me to follow, but then I don’t have a science background. I’m not aware of any other blogs which really attempt to explain the science from first principles in quite the same ways as SoD, although obviously Real Climate is also very educational, so I’m certainly grateful to Steve Carsen for his efforts.
I’m currently taking David Archer’s “Global Warming:The Science of Climate Change” course at Coursera which I’m finding really interesting.
speaking of content free response: I still haven’t seen any evidence whatsoever which could back up your baseless claims. Technobabble doesn’t count as evidence. Nothing more than an appeal to your own engineering “brilliance”. It reveals once again that you haven’t bothered to check the literature on the subject outside your field. Of course that doesn’t restrain you from never-ending innuendo. Bad faith of the worst kind indeed. Let’s face it, if I want to make anyone on Earth listen to what I have to say, the first thing I would make sure at all costs is that I treat the opposite in good faith. David Young, you are light years away from that.
It may well be the case that climate scientists might benefit in some cases from involving professional statisticians in their work. It may also be the case that statisticians would benefit from involving professional climate scientists when they are trying to analyse the results of climate science from a statistical perspective.
Andrew, indeed I think that’s probably true. My issue is with a simple statement like “giving statisticians a more prominent role will improve climate science”. There’s an element of truth, but it’s such a complex issue that blanket statements like that have little overall value.
This is an old contrarian meme from the smearing of Mann, the insinuation being that Mann had botched the statistical analysis used in MBH98/99. It has been repeated by contrarians as part of the Great Chorus of Fake Doubt ever since.
Actually, some very good statisticians have been involved with climate science. ANyone who thinks otherwise simply does not talk to the right people.
BY the way, Stockwell is a member of PSI, home of the Slayers, whose publisher Stairway is sponsoring Salby’s UK tour this week. There’s still time to sign up and report.
John, I’m sure you’re right. Something else that people often won’t acknowledge, is that many scientists are also very good statisticians. Just because some get their statistics wrong doesn’t mean that all scientists can’t do statistics.
Stockwell being a member of PSI is not that surprising given his publicly stated views. I had considered attending one of Salby’s talks, but I think that would be hard to do given that I’ve decided to remain anonymous. I don’t think the venues are big enough to hide in 🙂
@ John Mashey
Agreed re good statisticians involved in climate science – perhaps not fully clear from my previous remark. But mainly thanks for pointing out that Stockwell is a member of PSI. I had no idea. Another reason to increase the size of the pinch of salt.
Apologetics. Yeah, David, I get the message. The effect would have been better if you’d not painfully explained apologetics, though.
Everything’s information. In the case of David’s continued leakage it’s now fairly clear to me that he’s a waste of time. Discussion with David is like talking to a teenager; a simple request to clean the closet unleashes the infant barrister.
Indeed, many scientists are good statisticians, and where I first worked, were expected to consult great statisticians (like John Tukey, etc) as needed.
AS a more specific example, see Strange Scholarship, PDF pp.67-70, on an NCAR workshop in 2007 that combined climate scientists who are generally decent statisticians with statisticans who (mostly) had done some climate-related work. Of the 17 participants, I’ve spent some time in person with 4 in small groups, and corresponded with 6 more, some often.
“How can statisticians become involved?” is excerpted from (statistician) Jim Berger’s talk, recommended.
During the mid-2000s, climate shifted somewhat from physics arguments, often too easy to refute via conservation arguments or simple analogies, to statistics arguments, where:
a) Somebody could do a lot of statistical analysis that could not possibly be followed by the general audience, but were designed to create doubt.
b) They might well use cherry-picking, or odd methods, but could generate compelling-looking graphs that would take some effort to untangle.
c) Statistics easily can be mis-used to raise unrealistic doubts, but fortunately, most statisticians (and climate scientists) try hard to quantify uncertainty, never trivial.
Of course, the Slayers are still fighting physics, i.e., NO GREENHOUSE EFFECT.
I’m sure they are pleased to have Stockwell with them.
WOTTS: I wasn’t expecting you to attend Salby, but who knows, perhaps one of your readers might. There are 3 chances left, 2 in London, 1 in Edinburgh.
John, my response was partly in jest 🙂 Quite possibly, and if someone does it would be good to get some feedback.
@ John Mashey
Re: your (a)(b)(c) above – yes, yes and yes, but especially (a) and (b). It is quite possible that I acquired this fundamental insight into the evolution of denial from your writing in the first place. If it was from you, thanks again.