There is a provocative new paper available at an online discussion journal:
Makarieva, Gorshkov, Sheil, Nobre, Li: Where do winds come from? A new theory on how water vapor condensation influences atmospheric pressure and dynamics. link
See discussion at the Air Vent (here and here)
I first became aware of Makarieva’s research about a year ago, I encountered her at a blog (probably climateaudit) and suggested that she send me copies of her papers. My curiosity was struck initially by her hurricane papers. We began an extensive e-dialogue about her work, and I started to dig deeply into her hurricane papers. And then climategate struck, and my attention unfortunately became diverted. I am delighted to take this opportunity to revisit her work and discuss her new paper at Climate Etc. Anastassia Makarieva is one of the people that I have invited to host a thread at Climate Etc.
In my opinion, the most significant characteristic of an important paper is that it changes the way you think about a problem. This qualifies as an important paper, by my standards. While the paper is controversial, it has potentially far reaching implications for our understanding of how the climate system works.
Summary: This paper presents a theory as to how condensation influences atmospheric pressure through the mass removal of water from the gas phase with a simultaneous account of the latent heat release. The mechanism described by Makarieva et al. is correct, and it is not currently included in climate models. It is not clear to what extent this mechanism “matters;” their thermodynamic analysis is insufficient to demonstrate the relative magnitude of this effect. Nevertheless, this mechanism raises important issues regarding the structural adequacy of the atmospheric dynamical core used in climate models. Given the overwhelming importance of water vapor and cloud feedbacks in climate model simulations, re-examination of the atmospheric dynamical core used in climate models is called for.
Note: the summary paragraph will be revised based upon the discussion, as per TomFP’s suggestion.
“Abstract. Phase transitions of atmospheric water play a ubiquitous role in the Earth’s climate system, but their direct impact on atmospheric dynamics has escaped wide attention. Here we examine and advance a theory as to how condensation influences atmospheric pressure through the mass removal of water from the gas phase with a simultaneous account of the latent heat release. Building from fundamental physical principles we show that condensation is associated with a decline in air pressure in the lower atmosphere. This decline occurs up to a certain height, which ranges from 3 to 4 km for surface temperatures from 10 to 30 °C. We then estimate the horizontal pressure differences associated with water vapor condensation and find that these are comparable in magnitude with the pressure differences driving observed circulation patterns. The water vapor delivered to the atmosphere via evaporation represents a store of potential energy available to accelerate air and thus drive winds. Our estimates suggest that the global mean power at which this potential energy is released by condensation is around one per cent of the global solar power – this is similar to the known stationary dissipative power of general atmospheric circulation. We conclude that condensation and evaporation merit attention as major, if previously overlooked, factors in driving atmospheric dynamics.”
Review
The paper makes its arguments using the first and second laws of thermodynamics, conservation of mass (in the form of the mass continuity and scalar continuity equations), and the hydrostatic equations (my text Thermodynamics of Atmospheres and Oceans explains the basics behind these equations as applied to the atmosphere). Their analysis does not include the Navier-Stokes equations or a dynamical (time dependent model). Rather they calculate mechanical work derived from the atmospheric pressure gradient induced by condensation processes.
The basic points made in this paper are valid, and the mechanism they have identified is a real one. There is one point where I disagree with the authors, and that is on evaporation vs. condensation. They identify “salient differences” between them which in fact do not exist. Evaporation is not a surface specific process. When a cloud forms in the atmosphere, the condensed water has one of two fates: fallout in the form of precipitation or evaporation. The precipitation efficiency of clouds is rather low, much less than 10%. So most of the condensed water in the atmosphere eventually evaporates in the atmosphere. But I don’t see that this has much impact on their overall argument.
It is not clear to what extent their mechanism “matters;” their thermodynamic analysis is insufficient to demonstrate the relative magnitude of this effect. They provide some back-of-the-envelope estimates for the Hadley cell and hurricanes. Investigation of whether this matters requires a more comprehensive scale analysis that includes the Navier-Stokes equations and model simulations to test these ideas. The time and space scales of the adjustment to the small mass disturbances engendered by condensation, and how the adjustments occur in the vertical or horizontal direction, can only be determined in the context of a simulation.
I think that this mechanism might have some effect on storms (e.g. hurricanes and extratropical cyclones) where evaporation and precipitation are large. In terms of climate and larger scale circulations, I think the biggest effects would be in areas where there are large gradients of evaporation minus precipitation (see here). In Dec-Feb, large gradients are seen near the Atlantic coasts of South America and southern Africa, and between tropical and subtropical ocean regions. In June-Aug, large gradients are seen in the Australasian monsoon region, north-south gradients on the South America and African continents, and between the tropical and subtropical ocean regions. Climate models do not do a good job with monsoon circulations. Weather models do not do a good job with hurricane intensity. It is worth testing to see if their mechanism could possibly provide at least a partial explanation for these deficiencies.
Implications
The reason this research resonates with me is that I have long been concerned about the fitness for the climate task of the atmospheric dynamical cores designed for weather models. Yes, the use of the weather model cores provides confidence that they climate models can simulate weather systems. But what about processes involved in water vapor and cloud feedback, the big gorillas in driving the sensitivity to greenhouse warming?
So are climate models wrong? Well, there are a number of simplifications made in the dynamical core of the atmospheric models. These simplifications derive from the history of numerical weather prediction, and have arisen variously from considerations of numerical stability and computational efficiency to simplifications introduced over 50 years ago into moist thermodynamic equations for ease in deriving analytical solutions. While continuing the use the same atmospheric dynamical core used in weather models, climate models have evolved to have far more sophisticated radiative transfer and cloud parameterizations and most climate models now have prognostic equations for cloud water content.
While the current atmospheric dynamical cores provide mostly sensible simulations of the broad features of the atmospheric circulation, there are some nagging concerns. Is it possible that the impacts of all these approximations accumulate in the tropical upper troposphere, where there is some large discrepancies between the models and observations?
In What can we learn from climate models?, I quoted this statement from Thuburn (2008), which directly targets the issue raised by Makarieva et al.:
“Moist processes are strongly nonlinear and are likely to be particularly sensitive to imperfections in conservation of water. Thus there is a very strong argument for requiring a dynamical core to conserve mass of air, water, and long-lived tracers, particularly for climate simulation. Currently most if not all atmospheric models fail to make proper allowance for the change in mass of an air parcel when water vapour condenses and precipitates out. . . However, the approximation will not lead to a systematic long term drift in the atmospheric mass in climate simulations provided there is no long term drift in the mean water content of the atmosphere.”
In thinking about how the atmospheric dynamical core could be improved to address processes associated with water vapor and clouds, I am wondering whether climate models need to be nonhydrostatic and also might need to account for multicomponent fluids and multiphase flows, where moist air is treated as a binary fluid. Staniforth et al. describe the new nonhydrostatic formulation for the weather and climate models at the UK Met Office. The only application of multiphase fluid dynamics I have seen for the atmosphere is a paper by Bannon (note there is controversy over eq 5.16 in this paper). This subsequent paper by Bannon has ideas whereby the mass conservation can be incorporated without a full multiphase treatment. Note, I am not an expert on the nuances of the atmospheric dynamical cores, I am throwing these issues out for consideration.
The model that comes closest to addressing these issues is the WRF model (which is not a climate model). A relatively efficient way to proceed would be to work with WRF model to assess the impact of neglecting these various terms. Simulations of the impact of precipitation mass sink on hurricane intensity have been investigated using WRF, showing a modest but non-negligible impact (this is not published to my knowledge).
Now on the topic of thermodynamics, there are a whole host of mostly unnecessary simplifications that have been introduced that are probably pretty irrelevant on weather time scales but probably relevant on climate time scales. A topic for another post, but as an example, the heat capacity of condensed water is commonly ignored; this is ok in an analytical/theoretical analysis such as Makarieva et al., but not for a climate model.
Makarieva et al. also raise the issue of the nature of interconversions among kinetic energy, potential energy, and internal energy. To anyone that has a copy of my textbook, the relevant discussion is in section 12.2. I will do a post on this at some point, there are some issues that have been nagging me about how all this is treated, and I want to dig into these a bit more.
Addendum: the “community” response
If you go to Makarieva’s web site, you will see that she has published several papers on this general topic in journals that include the prestigious Proceedings of the Royal Society and Physics Letters. However, she has not succeeded in getting her work published in atmospheric science journals.
Her previous paper submitted to ACPD received a huge number of comments (by the standards of the journal), and the paper was eventually rejected. The discussion is well worth reading.
The previous paper received some discussion on the tropical listserv. Some people dismissed it since Emanuel’s MPI theory seems to work just fine. Others had misconceptions about what the paper said. A few people seemed quite interested.
Bottom line: it is challenging for an “outsider” to get a paper published that poses a major challenge to the status quo. Insiders are less likely to challenge the status quo, so outside challenges should be welcomed and considered carefully. Part of the challenge is for the outsider to spin up in the “culture” of the field and cite the relevant literature and use terminology and notation that is familiar to the target audience. And not to overstate the case. I think the present paper will have an easier time in the review process than its predecessor at ACPD, we shall see.
And it is important for “insiders” to engage with the “outsiders.” I understand that Makarieva has contacted a number of climate scientists for feedback on her papers; a few have helpfully replied. I guess my interaction with Makarieva “counts” as interacting with a skeptic, since I encountered her on a skeptical blog and she is challenging the status quo. If this is the kind of thing that I shouldn’t be doing according to the IPCC “in crowd” (see here), then the climate field is in a great deal of trouble. I personally think that we are very fortunate to have Anastassia Makarieva (an esteemed Russian biophysicist) and her colleagues taking a look at some of the fundamental assumptions made in our field. Whether all this turns out to be important or not, I don’t know. But I do know that it is important for us to try to figure this out.
Dr Curry:
Thank you for the quotation of the discussion between ‘TimTheTool Man’ and ‘gavin’.
There could not be a more clear example of lack of scientific rigour than the argument of ‘gavin’: the greatest error any scientist can make is to assume he/she is right so any doubt of his/her assessment should be ignored and not considered.
One can only hope that this behaviour of ‘gavin’ is rare in climate science but, sadly, my experience from interactions with other climate scientists indicates that it is not.
Richard
Richard, Richard, Richard-
Don’t you know who “Gavin” is?
You poor thing.
It is much worse than you think; much, much, worse.
You say: “So most of the condensed water in the atmosphere eventually evaporates in the atmosphere. ”
But pretty much the only way to get water out of the atmosphere is for it to condense. So while it may cycle around between gas and cloud several times, eventually it must end falling as precipitation.
So at first pass it shouldn’t affect the basic premise. Although it may account for local differences as the cycling between phases make take place as the wind pushes things from one area to another.
Your citation
Thuburn 2008
ends, ¨However, the approximation will not lead to a systematic long term drift in the atmospheric mass in climate simulations provided there is no long term drift in the mean water content of the atmosphere.”
re ¨…mean water content…¨
R Pielke Sr comments on fossil water studies , recently; have we brought into the atmosphere sufficient quantities of ancient water to produce a ¨long term drift?¨
ot – Was it Feynman who told us about heretics? SciAM article title honors you with the title.
Thanks for the solid contribution, John
A few people have commented on Tim’s approach. I thought this one about the role of models was pretty good.
” TimTheToolMan said: “How about you ’splain to me what that has to do with a discussion centred on the risk associated with using model results to build our science upon?”
There is absolutely nothing other than models upon which to base science.
All science is models. Indeed all human understanding is models. Hawking does a brilliant job explaining this in “The Grand Design” with Leonard Mlodinow, (Kindle location 411) . This is a brief snippet. Hawking discusses this at some length. Models are in no way unique to climate science.
“There is no picture- or theory-independent concept of reality. Instead we will adopt a view that we will call model-dependent realism: the idea that a physical theory or world picture is a model (generally of a mathematical nature) and a set of rules that connect the elements of the model to observations. This provides a framework with which to interpret modern science. “
“-All science is models.-” I like that. (The fact that it refers to Hawking is a bonus.) Whether it’s quill pen and parchment or the latest computer gadgetry, it’s all down to models.
I’m also rather taken with the “All models are wrong, but some are useful” quote from the statistician, George Box. It shifts the argument. If a model is useful then the possibility / certainty that some part of it may be wrong becomes a minor quibble rather than a major challenge.
I confess I’ve not looked at Makarieva yet. Will do so later.
Dr Curry,
“But not to worry: RealClimate says the models are fine”
This seems like an unreasonable and antagonistic portrayal of the viewpoint expressed expressed at RealClimate. You’re discussing the details of a particular paper discussing a specific issue and then linking it to a thread elsewhere where a poster is saying “So hey what if it was discovered there was something wrong with the models what would that mean for all the research based on them and the entire field of climate science?”
The implication of “TimTheToolMan”‘s line of question was obvious: If some unnamed deficiency could be found in the models that would somehow completely invalidate them and any work based on them. The response given at RC was fairly clear: The models are known to work where they’ve been demonstrated to work and you can’t invalidate that by finding something for which they don’t work.
Actually I thought the Makarieva paper was a perfect example of what TimTheToolMan was talking about, and the Makarieva paper was even brought up on the thread.
I understand that however my issue is that you’re presenting a very specific case and then presenting the realclimate view of that case as “But not to worry: RealClimate says the models are fine”.
In actuality realclimate are commenting on something else i.e. an unnamed vague inference of a potential problem with the models and in any event didn’t say anything close to “the models are fine”.
I’m nowhere close to having the skill and knowledge necessary to evaluate either the Makarieva paper or to judge its implications for climate science. I can, however, read plain English just fine and I can see the manner in which you brought up realclimate is disingenuous and your summary of the expressed position is erroneous.
How therefore can I have confidence in your reading of the Makarieva paper and the implications for climate modeling? I am reliant on experts such as yourself to evaluate these things and tell me what they mean (if anything). If an expert shows their judgement to be unreliable then it reduces confidence in it.
Well you are certainly entitled to your interpretation and to your criteria for choosing experts. I am not asking people to trust me, I am presenting arguments for them to consider. I’ve decided to change the title of the subheading to Comments from RealClimate on Models.
Then why not include Gavin’s comments at tAV where he engages with one of the authors and not just the comments he made to TimTheToolMan on RC? His comments at tAV are much more technical and informative that the ones you chose to present.
I find it curious that Dr Curry did not include this part of the exchange:
“20Marco says:
22 October 2010 at 1:00 AM
Gavin,
Tim refers to the new meme in parts of the deniosphere. See e.g. Jeff Id’s blog:
http://noconsensus.wordpress.com/2010/10/19/momentary-lapse-of-reason/
No need to post this.
[Response: I know what he is referring to, I’m ‘Gavin’ on that thread. But what I object to are coy hints of some looming issue that everyone is running scared from. That is simply spin for the unwary. This is not as exciting as the authors claim, and the response to their earlier claims is full of very good reasons why. – gavin]”
http://www.realclimate.org/index.php/archives/2010/10/unforced-variations-3-2/
Presumably she read it, as it was posted half an hour before a comment that she did quote. If you follow the link to AirVent your will find that the “climate communities” (Gavin’s) responces are cogent and to the point.
In other words, contrary to the impression Dr Curry appears to trying to cultivate, climate scientists respond to detailed scientific critics with detailed analysis, and where necessary, corrections. In this case, according to Gavin, the magnitude of the effect is very small (0.6% of pressure), and models were already being modified to include the effect before it was pointed out by Makarieava.
It is true that Gavin give flyby sceptics making vague and unsubstantiated points short shrift on Real Climate. I cannot blame him, having myself given detailed and devestating rebutals to such posts only to be ignored (both in print and in opinion).
I think this is a very good example showing that many people do not understand what our work is about when trying to judge about the magnitude of its implications. Dr. Schmidt’s words about 0.6% of total pressure refer to the weight of vapor molecules in an atmospheric column. This is a store. The effect of condensation on pressure gradient that we are describing depends on flux — the rate of precipitation. The faster the vapor is removed from the atmosphere, the greater the resulting pressure gradient will be. Fluxes and stores should not be confused, as these are variables of different dimensions.
I would also like to encourage people, before spreading the word about 0.6% of total pressure as a negligibel magnitude, to give a thought to the relevance of this magnitude in the atmospheric context. For example, Hadley cells (the major feature of Earth’s atmospheric circulation) are driven by a pressure difference that is of the order of 10 hPa, i.e. a meager 1% of total air pressure.
This is Real Classy Tom.
Tom: “Tim refers to the new meme in parts of the deniosphere. ”
I went to airvent and searched for one comment by Gavin. It seems to be in reference to surface pressure being the sum of the mass in a column of air directly above, which is true but completely misses the point.
Makarieava talks about differential pressures in the atmosphere being the driving force of atmospheric dynamics, and how the differential pressure could be caused by precipitation, evaporation and the energies associated with this process. She explains how the models used today fiddle with the constant values to get the right shape and size of atmospheric dynamics.
There is probably more to this discussion, but i didn’t bother reading anymore because they were just talking about different topics. When having a scientific discussion the first point would be to agree on the topic being discussed.
First, if you are going to quote me, quote me. The words you quote were not mine, but those of Marco whose post I quoted.
Second, on the Airvent, Gavin posted more than six posts on the topic. In some of them he was explicitly trying to clarrify the nature of the thesis; and he gave cogent reasons to think the effect is minor either as interpretation of the effect of precipitation on column air pressure (the mass of H2O in the total air column represents only 0.6% of the total mass so the effect is inconsequential) and the local pressure changes due to condensation (if the effect is large, when new clouds formed by condensation, they would compact due to the effects of pressure equalization; and, the far more significant effect of condensation is the release of latent heat which swamps in magnitude the change in pressure due to condensation). I do not know that Gavin is correct in these points, but they are certainly relevant and plausible (ie, cogent). In going outside of his own blog to discuss the topic in this way, he is doing just what many are claiming he avoids at all costs.
Perhaps the first point you should be considering is that when considering a scientific discussion, you should consider the entire exchange.
I do not know much of what other people are claiming, but, in my opinion, the comments of Dr. Schmidt on the Air Vent, although I disagree with most of their scientific content, were reasonable and constructive. This attention is certainly much appreciated by the authors, as I expressed there.
This is one of the few times that Gavin has posted on a sceptic site and actually discussed the issue in a rational manner. Trying to use this example to imply he does this regularly…
You are right Tom, not your statement. I apologize.
Yes, Gavin wrote several posts. I only read the first which was on a different topic from the essence of the paper IMO, the main topic in my post.
I think Anastassia’s results are about the science, not the models. Your kind of science, where she says something major has been overlooked. Do you think she’s right?
In particular, can you see where her Eq 34 comes from?
Re eq 34. As far as i can tell, she defines S a sink/source term for water vapor in the context of saturated adiabatic ascent. It is difficult to figure out exactly what this is since she uses non standard notation. This is what i meant in my original post about it being easier to gain acceptance (and understanding) if the conventional variables and notation are used. S should be the rate of condensation, presumably in saturated adiabatic ascent, which is not entirely clear from the equation as it is written. It would have been better to incorporate eq 6.41 from my text to clarify that this is actually saturated adiabatic ascent. (if you don’t have my text, email me and i will send you ch 6)
Anastassia believes she has calculated her hadley cell example using eq37 entirely independently from any latent heat. I’m not sure that is true considering that observed values were used. Basically she is making the point that the latent heat energy is so inefficiently converted that it is a negligible factor.
My estimates show latent energy is about 5X larger than that caused by condensation volume loss.
I’m not sure what to make of that but the key is in the hadley cell example.
Good point, Jeff.
Consider one man heating an air column in atmosphere A, another man removing gas from an air column in atmosphere B. Measured in J/sec, power that man A is exerting is five times the power of man B.
What will be the pressure gradients generated by the two men in the two atmospheres? We calculate this for man B, and show that these pressure gradients are just what is observed.
While man A exerts a higher power, he may not generate any gradient at all. This is because all this power will leak through heat conductivity. To know the pressure gradient generated by man A, one has to theoretically derive the efficiency of this thermal process. This has never been achieved, such that this efficiency for atmospheric processes is simply tuned to observations. So far there has been no any candidate to drive winds, so there was simply nothing doing.
In comparison to man A, the power exerted by man B is pure work, which converts to kinetic energy with 100% efficiency.
So, one should be careful about comparing joules. There are joules and joules, this is what all the thermodynamics is about .
My phrase above “So far there has been no any candidate to drive winds, so there was simply nothing doing.”
should have read
“As long as there has been no any other candidate to drive winds, there was simply nothing else doing.”
my apologies for language errors.
Yes, Eq. 34 contains the same physics as Eq. 6.41. Eq. 6.41 is key. It will take me some time to show it using symbols and formulae in a blog comment, but I will do so today or tomorrow. If it is possible that the chapter be emailed now, I would greatly appreciate, as I do not have all texts at hand right now.
Anastassia (and Judith)
6.41 is an enthalpy equation, incorporating L and cp. It’s basically a heat budget. Eq 34 doesn’t involve these quantities, and seems only explicable to me in terms of mass conservation. Presumably combining 6.41 with mass conservation would be useful.
Judith,
Thanks for sending that chapter. As I understand 6.41 is an enthalpy equation giving the rate of condensation in saturated adiabatic ascent. I didn’t see that Eq 34 specified saturated conditions?
My interpretation of their narrative is that what they intended. “S (Eq. 34) is the sink term describing the non-conservation of
the condensable component (water vapor).”
Nick, the line right below Eq. 34 specifies saturated conditions, see line 12 on page 24030.
Not that I can see – that line simply defines, as a matter of notation, Nv as the density of saturated water vapor. An explicit statement would help.
I am not sure I understand what you mean. Nv is defined as saturated molar density. Eq. 3 specifies how saturated pressure depends on temperature. Dependence between density and pressure is set by ideal gas law, as indicated prior to Eq. 37. Temperature depends on height as moist adiabat. What is missing?
Not in response to any comment except Nick’s. The model sidetrack was my own, or our own rather. Anastassia’s paper stands by itself. I was unaware that the models didn’t include the mass effects and when Nick linked to the parametrization of a hadley cell, we found that the volume loss wasn’t accounted for. Gavin chimed in with some very helpful comments, and although he blew the heat calculation as far as I can tell, his point was well taken.
“I’ve decided to change the title of the subheading to Comments from RealClimate on Models.”
That at least helps to solve the misrepresentation/antagonism problem.
However I (and the reader) don’t know the relevance of these comments to the rest of your post or what it is you think about then. We can reasonably conclude you hold a negative opinion but have no idea what it is or why.
Comments #1 and #2 are in reference to the vagueness of what “TimTheToolMan” was talking about so I don’t understand their inclusion here at all concerning a specific issue.
The point in comment #3 seems reasonable: Models work where they’ve been shown to work. Ideally Dr Curry you would show how the Makarieva paper impacts models in areas they were previously thought to work and reveals how erroneous conclusions were reached due to a faulty understanding of climate behavior.
Comment #4 seems close to an areas you’ve discussed yourself i.e. confidence in models and that we neither need to believe models are perfect or that models are useless but rather we can use them where they’re known to be good and not in other cases.
My sympathies are with Gavin in that exchange. I’m sure he gets bombarded with this stuff = what if you were wrong = wouldn’t this turn everything upside down?
It’s lazy stuff, and you can do it for anything. Of course if you were wrong, something would change. That’s why you try to be right. How many ways can you say it?
ToolMan had nothing substantive to say – just noise.
Don’t know which particular comments led to this, but here’s TimToolman’s latest contribution.
“Thank you all for your responses to my question regarding the implications of a significant model flaw. Some of the answers were very enlightening and I appreciate your time in considering this as you have.”
So maybe the fuss and bother was worth it.
Anastassia says
“We conclude that condensation and evaporation merit attention as major, if previously overlooked, factors in driving atmospheric dynamics.”
Do you think they are previously overlooked?
Yes I do, by virtue of not including source and sink terms for water vapor in the mass continuity equation. Whether or not this is “major” is yet to be established.
Prof. Curry,
(Sorry if it is a bit of topic)
I’m a skeptic and I have no respect for Michael Mann or Gavin Schmidt. They seem to believe that they are to smart for the rest of the commoner or even other scientist who dare to not agree with their believed infinite wisdom. I would say that they are the reason why I am so cynic toward climate scientist.
When I first encounter you and your work it was when you publish your paper on hurricane a few years back. At first you reacted to criticism in a similar way that Mann and al. did. Then you invited Steve McIntyre to your university, it took me by surprise, but it also helped change my opinion toward you. Since then you have engaged with skeptic and even if I don’t always agree with your conclusion I do have much more respect with you.
To be More on topic. Climate science is a young science by any standard, and that even though if human have been worried by climate since the antiquity. For a long the gods were behind the atmospheric phenomenon. It is only a 100 years since Aehrrenius established the logarithmic effect of co2. And the first computer model which applied this effect are not much more than 30 year old.
Knowing that, and considering our any other science have had huge evolution that metamorphosed our understanding. It is clear that it is very presumptuous for climate scientist “à la Mann et al.” to believe that their understanding of the atmosphere will never be revised in any way. Or in other term, their knowledge is finite and can only be slightly improved.
Paper like this one, or the cosmic rays theories, or Spencer theories, or Lindzen’s theories provide many example of how it can be different. Each theory has its caveat and its good point and only better data can help clarify was is the best way to go. The problem is with the massive amount of money invested in computer model, it leaves not much for improvement of data quality, thus improvement of climate science.
Dr Curry,
Thanks for my 5 minutes of fame :-)
I was also very interested in the responses of those in the thread and both Jim and Gavin were very defensive of the models and neither would entertain the idea there might ever be a serious problem with them let alone be prepared to discuss any possible risk mitigation for the future.
It was also interesting to note that the only other reply from Gavin came to another poster Dan H. who tried to give his view on what might happen… His idea was a good one in that essentially the political backlash may well outweigh the scientific implications.
Gavin again disputed the possibility again calling it a “fantasy land” and ” a recipe for unhinged speculation that is completely pointless” Ironically Dan H ended off with “We will see if people respond with enthusiasm or meanness.”
I’m like sharperoo in lacking the skill and knowledge to deal with the details of Makarieva’s paper but I found it disappointing – no mention of monsoons. (Though there was one title in the referenced papers.)
John’s quote ” have we brought into the atmosphere sufficient quantities of ancient water to produce a ¨long term drift?” reminded me of a thought I had about the amount of water that’s been held back from river and ocean circulation by dams and other water storage / diversion structures. (Probably someone else’s remark about rate of SLR being artificially constrained by this – though my thought would be that’s probably not enough water to affect SLR.) Back to the quote, I’d suspect that the mining of the Artesian basin and Indian and American aquifers contribution to atmospheric circulation would be offset to some measureable extent by other activities which allow more water to percolate back through the soil. Just not enough, fast enough to replenish the groundwaters.
That’s vintage Gavin Schmidt. That’s the way he responds to anyone not appropriately subservient to his viewpoint, no matter how civil and rational his opponent may be.
Well, RealClimate is Schmidt’s blog and he can handle it as he pleases. But he is mistaken if he imagines his haughtiness — to use the word from SciAm for how climate scientists can appear to the public — doesn’t detract from his cause.
Climate scientists can manage their patch of academia as their own little fief and they can abuse outsiders as they please. I don’t like it much. I see no cause for such rudeness; it seems far more adolescent than scientific. But it is their patch.
However, these scientists are mistaken if they believe the rest of us citizens and taxpayers owe them any support or allegiance when they come to us for help in saving the world from a crisis that gives climate scientists even more power.
These scientists already sound drunk with power and they already have been caught abusing that power. The rest of us may not be able to follow the arguments, but we know arrogance when we see it or hear it, and we know that arrogance often makes for poor, self-serving judgment.
Earlier John R T mentioned the late, great Richard Feynman. He could be tart at times, but I can’t imagine Feynman responding in public as Schmidt routinely does.
I don’t find the moderators on RC terribly rude – perhaps I spend too much time with engineers and mathematicians. They’re all pretty brusque with each other or scornful when errors are under discussion. I suppose it might seem high-handed or dismissive to a casual observer, but I have noticed they’re pretty supportive of high school students who might try to join the adult conversation.
“I don’t find the moderators on RC terribly rude ”
Previously I’d only heard stories of how RC moderated out people’s point of view. Now I have first hand experience of how a perfectly cival post exploring an issue they didn’t like simply got dumped.
Oh really.
I can’t tell you the number of times my comments have not made it through moderation.
Especially when I support my comments with peer-reviewed links that are contrary to the political agenda over at Real Climate.
The typical ploy they use is that a comment is allowed, with a snappy, brusque and off target response by a moderator.
Then a third party comes in and makes a nasty attack on the hapless, helpless truth sayer.
The final kick is a fourth person agreeing with the third.
Any responses by the original poster are blocked.
It has happened to me several times; earlier this year.
Less recently now that I have sussed out their modus operandi.
Re: some blog’s comments policy —
(1) In the case under discussion, you already know their aggressive stance.
(2) Copy, paste, and save locally before submitting.
(3) If the comment fails moderation, remember that it is the blog owners’ printing press. Submit the copy to a more sympathetic venue covering the same topic.
(4) If you feel strongly, start your own blog. They are free.
hixley, you seem to be inhabiting an alternative universe.
The response at RC is generally very good – more so if you consider that gavin et al are not casual bloggers, but active research scientists doing this in addition to their other work.
It never ceaes to amaze me, the time and detail they (Gavin especially) go to in responding, with links, direction to references etc.
What does get short shrift, is repeated posts that ignore the information provided, or make broad and unsupported allegations of malfeasence, or deny basic physics and persist in such despite correction.
Bad faith in argument deservedly gets some rudeness. IMHO they often get less than they deserve.
Michael:
You assert:
“The response at RC is generally very good – more so if you consider that gavin et al are not casual bloggers, but active research scientists doing this in addition to their other work.”
And your point is?
They chose to set up their blog as an organ (funded by Soros) to promote their work. They can – and do – operate it as they see fit. Similarly, Dr Curry has established her blog.
But Dr Curry encourages scientific discussion on her blog while RC rejects scientific discussion. This rejection of scientific consideration by RC is clearly demonstrated by the provided example.
‘TimTheToolMan’ asked;
“If the models were to be shown to be specifically deficient in some area and need significant rework what impact would you see that having on the thousands of papers that have relied on them to this point and of climate science in general?”
There are only two scientific responses to that question.
At one scientific viewpoint, the answer is:
“I do not think there is a real possibility that the models could be in error to a degree that would render their indications sufficiently erroneous as to affect the general implications of climate science. However, if that were the case, then … ” and, perhaps “but there is not sufficient space to discuss that here”.
The other scientific viewpoint would give this answer;
“Of course, there is always the possibility that new information can overthrow any presently solid scientific understanding (ignoring that possibility is known as the Kelvin Fallacy). So, although your question is hypothetical, if such new information were to arise then …”and, perhaps “but there is not sufficient space to discuss that here”.
But the answer from ‘gavin’ was;
[Response: And if the moon were made of green cheese, what impact would that have on space science in general and on the astronauts who walked on it? Please don’t play games. – gavin]
And this ‘gavin’ says he is a scientist!
Richard
Well typed, Richard. Very well typed.
In fact, to the question asked by “TintheToolMan”, the only legitimate responce is that “The impact on scientific papers will entirely depend on the nature of the discovered deficiency; and as that has not been specified, no intelligent discussion of the potential impacts can be made.” We do not know from TintheToolMan’s vague hypothesis whether the supposed deficiency would result in an overestimation or an underestimation of warming, or have any impact at all. How then can we concievably discuss the impact on particular (unspecified) papers?
What TintheToolMan was clearly fishing for was a responce along the lines of “well of course those papers results would be in question”, where upon he would have sprung the Makariava paper on Gavin, and used his sound bite to focus discussion not on the merits of Makariava’s paper but on the supposed deficiencies of climate science. His question, in other words, was a rhetorical device to mount an attack on climate science without having to defend the merits of the Makariava paper, or even to show that it woud have the consequences for climate science that he wished to suggest.
The rhetorical device were treated by Gavin with the contempt that it deserved. Certainly he was not naive enough to fall for TintheToolMan’s rhetorical trap. This does not show that he was not interested in debating the scientific merits of the Makariava paper. On the contrary, he has already been doing so on The Airvent (first post: 12 noon, Oct 20th) before TintheToolMan made his first post (Topic opened Oct 21st).
In fact, TimtheToolMan had read Gavin’s comments on the Airvent, and even responded to them (post 131) hours before he could possibly have first posted his question on RealClimate. That demonstrates incontrovertibly that his question at RealClimate was an attempted rhetorical setup. He knew Gavin was discussing the paper calmly and in detail before he ever asked his loaded question on RC.
All of this begs the question as to why you and Dr Curry insist on focussing on Gavin’s responce to a rhetorically loaded question rather than his responce to a genuine discussion on the AirVent. Is it because only by considering exchanges out of context that they are conducive to character assisination?
If not, you and Dr Curry having been apprised of the context of the discussion will withdraw your innuendo and apologise for it. If you do not, it is your, not Gavin’s character that has been shown to be tawdry by this incident.
An interesting theory Tom.
Completely wrong, however. I was genuinely interested in having a discussion on the implications of flawed models.
If I’d wanted to “attack” Gavin, dont you think I might have used the Makariava paper directly? Instead I avoided it because I didn’t want to get into the specifics of flaws but instead “use the paper” as an indication that the discussion was warranted. I knew Gavin had knowledge of the paper and so I thought that the discussion would automatically have merrit. How wrong I was.
The responses I got were entirely unexpected.
Allowing on your claim that your intentions where genuine, your question was still poorly phrased. It was not capable of being the basis of sensible discussion. That being the case, why did you not take the opportunity to rephrase the question in a way capable of discussion?
We know why Tom.
Certain types visit Real Climate so they can pester, annoy and provoke. Not to learn mind you.
No, they just enjoy the experience.
The only thing more fun than trying to provoke a reaction, is getting one and then whining loudly about how mean and rude they are over at RC.
JC: lets drop the RC comments, not germane or moving the discussion forward
I did Tom. A number of times. Perhaps you missed that both Gavin and Jim point blank refused to accept that there might even be a significant flaw in the models. Once that was established it was impossible to progress to how one might minimise the impact of such a flaw…
In the light of their clear stance on this issue, the impact from the Makariava paper (if its non-trivial which on the face of it seems likely) will be even more interesting.
This scenario is chess, not science.
Richard,
Howapt that you would pick this particular comment.
This was just the kind of bad faith arguing I was talking about.
It was clear that ‘Timthetool’ wasn’t interested in a discussion – he was looking for some kind of silly ‘gotcha’, all too common amongst the keyboard warriors.
So what, if it was a “gotcha”? If it’s such an obvious “gotcha”, as you say it is, then – to coin a popular phrase – “it’s worse than we thought!” Climate science should not be so vulnerable to obvious gotchas, but it is, isn’t it? And therein lies the problem.
The appropriate response to Tim’s question would have been ” if new understanding undermined the method or function of climate models, then yes it reasonably COULD call into question the findings of papers whose conclusions depended on model output.”
The reason for this is, as any computer programmer outside the field of climate modelling (and perhaps some honest ones within it) will tell you, a computer model does not generate data on which conclusions can be drawn. A computer model is a model OF a hypothesis, not observed evidence thereon, nor a tested theory thereof.
But to recognise and acknowledge this basic inherent fallibility in computer modelling is to concede too much, no matter how rational or reasonable the question – and let’s not pretend that the question is not valid or is not a fair question, because it most certainly is, even if it’s packaged as a “gotcha” – because far too much of the current “state of knowledge” of climate science is not actually “knowledge” in a scientific sense, it is instead “virtual knowledge”, existing solely and wholly within the confines of a computer model.
And Gavin would rather propose that the moon is made of Stilton than publicly acknowledge the ramifications implicit in building upon foundations established upon climate models which it later transpired were wrong or deficient.
Simon, I would usually be the last person to go in to bat for Gavin. I agree that his comments on RC are often high handed and sometimes downright rude.
However, in this case I think people are commenting without thinking this through. If this was the only exchange between TimThe ToolMan and Gavin, then Gavin’s response would seem to be over the top. However, TTTM might be a frequent and somewhat vexatious commenter at RC. If this exchange has some “history” behind it, it would be more understandable.
Alex, what you say may indeed be fair and true. My experience, though, is that it’s actually quite difficult to be a frequent and vexatious commenter at RC. If one’s questions are searching or inconvenient, one’s ability to be a frequent poster becomes severely hampered. But I accept that TTTM’s treatment may have been different from mine and those of many others. TTTM’s question, whether he’s annoying or not, has merit.
Yes. That is my experience also. I have asked fair and honest questions about things that seem to conflict with the stream of the conversation. Often the questions do not get through moderation.
“However, TTTM might be a frequent and somewhat vexatious commenter at RC.”
I’m not. Well that depends on what you mean by “frequent”. I’ve certainly posted there a few times but not many and I think always in the manner of the recent post.
So that you can make your own mind up…
I had posted twice in the previous thread (and hardly at all prior to that) here at comment 147 and 149.
TTTM, I accept that you are not being vexatious at RC. Thanks for taking the trouble to clear that up.
“The reason for this is, as any computer programmer outside the field of climate modelling (and perhaps some honest ones within it) will tell you, a computer model does not generate data on which conclusions can be drawn. “
I don’t know where you’re getting the idea that climate scientists generally present this any other way. The realclimate thread discussed the problem of matching observation to model outputs and the entire point was that where observations have been shown to match models new observations can’t discount them.
I see a lot of stuff along the lines of “Oh if they were honest they’d admit that…” and yet when I look at what scientists are actually saying they appear to say exactly that. The “models are perfect!” type statements only ever seem to appear as strawmen erected by skeptics so they can then show some flaw in the models.
‘Hey, what if all of X is wrong?!’, is no ground breaking question.
Just hypothetical silliness.
“any computer programmer .. will tell you, a computer model does not generate data on which conclusions can be drawn.”
This statement disagrees with almost everything I know of computer programmers.
For one thing, getting all computer programmers to agree on anything approaches an impossible exercise.
.. or of computer programs.
Often faulty models will collapse on themselves, generating data that leaves small doubt the model is wrong.
Some models will collapse into lower order outputs, indicating the model is overly complex and leading to valuable conclusions about predictability and probability which were previously impossible.
One model, I’ve heard, expanded into higher order outputs through happenstance, and lead to the discovery of what is called the Butterfly Effect, which was quite the conclusion to draw.
Perhaps you mean that models of systems in dynamic equilibrium have no accurate predictive power beyond a certain period?
Which again is irrelevant.
There are many pragmatic, valuable, meaningful, provable applications of computer models that are simply not susceptible to redaction by counterexample.
As for the fraction of models (and their applications) that are so vulnerable, it ought be more – not less – easy to identify these automatable, machine logic cases than other types of error often introduced into science by human agency, and it would be a type of human error to not recognize dependence on the model in drawing conclusions that would become invalid because of a new understanding of some part of the model.
Not to belittle the value of scientific rigor, but scientific literature is rife with examples — from the Logarithmic Table with inverted digits not discovered for a century, to the Statistics Table likewise copied and used hundreds of thousands of time with the wrong results in one line to Leavitt’s Rule of 48 – where widespread mathematical errors have been found and fixed and the world did not end. (Well, okay, a bridge collapsed, but that involved multiple other failures of engineering.)
This seems much more likely to be a minor problem, if it ever happens at all, or a nonissue easily rectified, than a dagger to the heart of the life’s work of, well, anyone.
And if it isn’t, well done to Makarieva et al.
Michael, yours is the perfect example of how RC works. You respond to a person stating their post uses peer reviewed literature for support yet go off on a tangent about repeated posts ignoring the information provided…
Exactly what don’t you understand about RC ignoring wide swaths of peer reviewed science because it doesn’t match their bias??
My colleagues and I are very grateful to Dr. Curry for her genuine interest in the science we propose and for the many critical and specific comments she made on our work. I could write a very long post what I personally think of Judy and her role, “a portrait of Judith Curry by an outsider”, but at this first large-scale introduction of our work at Climate Etc. I should concentrate on science.
In our work we derive pressure gradients associated with removal of vapor from the atmosphere through condensation/precipitation. Nothing except basic physical considerations are used. On the basis of our derivations, we show that on the scale of Hadley cell, the mass removal process produces just the pressure difference that is observed (10 hPa). The same process also quantitatively describes hurricanes and tornadoes, as shown in greater details in our previous works.
Judy refers to these estimates as “back-on-the-envelope”. This raises an interesting methodological question. In physics, you cannot solve a problem without such basic estimates that give you the overall magnitude of the effect you want to study. Once such an estimate is obtained, it signals you whether to study the effect further or not. For example, had we concluded that the pressure difference would be 0.01 hPa over 2000 km, we would not be interested in talking about this further. At present we do not have any doubts that the effect is as significant as we propose.
In this sense the arguments of the Editors of our previous ACPD submission were logically correct. They attempted to dismiss our work at the same level of basic physical arguments. Dr. Rosenfeld provided a basic (Judy’s back-on-the-envelope) calculation proving that condensation cannot reduce air pressure. The Editors agreed, and the paper was rejected.
OK, now we show that condensation does lead to pressure drop. What is the next basic objection? No objections at this level any longer, as I can see. The argument has shifted to “the models that do account for precipitation, show that the effect is small”. In other words, the models are complex and account for everything, reviewed by thousands, and we trust them, not you.
However, science is not a matter of trust. Our physics is so transparent that for people who have built GCMs and described the climate system in all its integrity, it should be easily possible to come forward and say: hey, Makarieva et al., where you say it is 10 hPa, it is in fact 0.01 hPa, because in your derivations you did … instead of …. If you did … which is correct, you would obtain 0.01 hPa.
In other words, one can summarize how the effect is included into GCMs at the “back-on-the-envelope” level and contrast this derivation with ours. That would be alternative basic physics approach and people could see which of the two is wrong.
Thank you for your clarifying commentary.
Nastya,
First of all, I wish to express my admiration for the remarkably cool manner in which you conduct this discussion. It is obvious that you believe in your stuff so the patience that you show to the detractors is quite amazing.
Back to substance, it sounds like you are claiming to have refuted Rosenfeld’s calculation but I don’t see where.
One way or another, but the sheer difficulty of explaining your theory indicates that back-of-the-envelope calculations go only so far. If you want to prove yourself right your only choice is to hook up with some GCM (or weather prediction) guys and have them include the effect into a model. It cannot be too hard. Then we’ll know for a fact.
Even though I am on the skeptical side, I wish you good luck.
Judith,
This theory means “STOP THE PLANET”, no rotation required.
I’m one of the authors. Thanks for giving our text this attention. I wanted to put a few things in context.
The reason I got involved was that I could see that the ideas Anastassia and Victor were putting forward were not getting the constructive attention that they deserved. Some of that is doubtless due to the issues JC noted.
Last year, April, Daniel Murdiyarso (a climate scientist) and myself (a forest biologist) published an overview of the basic ideas for a less-technical audience “How forests attract rain: an examination of a new hypothesis” in Bioscience. That got some media coverage: you may have seen some of it: e.g. Mongabay, New Scientist and Scientific American (please google). I’m happy to share the PDF if anyone wants it. The point is that many of the wider implications (including monsoons) are considered in a reasonably non-technical manner for those of you who might be interested in that.
But what about that important technical audience? Anastassia, Victor, Larry have other technical papers for those who want to read up on that. You can lean more if you visit their site. There are a few critical reviews and commentries out there too. We should certainly welcome these – careful critical scrutiny is how science moves forward. But misreadings and assumptions have been getting in the way. So we have been working together to communicate the basic physical concepts as clearly as we can. So that is the reason for this new paper. It is certainly not perfect. We welcome feedback on how we might improve it.
Please note that our article is about basic physical principles (and is already quite long). It is not about climate models or even about most of the theory’s implications. If the principles are widely accepted then we can all look forward to examining and debating the wider implications (exciting stuff!) … but I suggest we do this step-by-step. For now,we invite careful critical scrutiny of the theory (is it correct? what is unclear? what are the problems? how can we test it?). Many thanks for all suggestions and comments.
Douglas
Douglas, you write ” how can we test it?” Surely this ought to be the key question that people should concentrate on. It is my pet thesis that the ONLY thing that non-validated models are any use for, is designing the next experiment. So, please, yes. How do we do an experiment to test this new work?
Makarievas Theory is far better then most as she follows the low pressure up into the atmosphere. There is a vast amount of strong heavy wind energy that then can be pulled to the surface to fill in the low pressure trail.
The density of the gases in wind higher up is far less then the heavier gases closer to the planet surface. But the mix then makes a far more dense wind pattern.
Makarieva just needs to consider the height of the low pressure bubble. The height of the monsoon low pressure bubble would be far closer to the planets surface than a hurricane low pressure bubble.
*blink*
Wow.
Is the “have-you-stopped-beating-your-wife” gotcha style of interrogation still so acceptable?
If anyone can explain the value Tim adds to the discussion with this abortive exchange, or the value of this digression to contemplation of the provocative Makarieva paper, it would go a long way to reconciling for me the stated goals set out in the welcome to Climate Etc with what is presented above.
Wouldn’t it be better to ask something like, “For what parameters or conditions is this Makarieva effect dominant or important?” or, “Is the order of magnitude of the ‘Makarieva effect’ large enough to affect outcomes in an important way?” or, “Is this actually a new effect, or simply something previously known and taken into account by a different name looked at with fresh eyes?” or “Does parity or symmetry of the effect or associated countereffects cause it to cancel out some, much or most of its impacts?” (or any of countless questions better informed by actual expertise in the topic.. which I certainly lack.)
Yes, it seems mean to try to knock down a new theory like some schoolyard bully, and likely most such challenges have already been addressed or could quickly be disposed of, but these are at least questions going to the merit of the ideas instead of the character of the author, and for that reason they’ve all got to be better questions than, “why are there scientists who get frustrated by transparent attempts to set them up for sophist’s tricks and traps?”
Scientists are human, sometimes even clever, hard-working humans who have devoted decades of their life at great personal sacrifice for the sake of advancing their field of study even a tiny measure.
When these humans attempt also that second profession of communicator and fall a little short in the area of courtesy — which in my experience of scientists is far less common than for the general population — I’m inclined to cut them some slack until they fall short of courtesy to someone honestly and sincerely seeking to better understand.
Which I frankly don’t have trouble determining long before the scientist’s frustration becomes apparent.
Judy
With the benefit of hindsight it may have been better to leave references to RealClimate out of your post. We all know that RealClimate is more about the politics than the science. The comments here have all been about personalities which does not really move the job forward. You need to rise above. Stick to the science. IMHO
Kind regards Gary
In hindsight perhaps yes, but the exchange does make a point that I think is important, I will try to clarify this in the main post.
Well, it seems the discussion is being sidetracked by the comments i pulled from the RC thread, which is unfortunate. The point I was trying to make about was about how the “community” responds to challenges to the status quo, particularly from an outsider. The issue Tim raises is an important one, really. How receptive is the climate community to the idea that there might be something structurally wrong with climate models, and how open are they to ideas from outsiders about different ideas on how to do this?
Personally I think it would be exciting if somebody came up with a new idea that would improve the structural form of climate models. Yes, much work is being done on improving parameterizations and adding new elements of complexity like carbon cycle. But not much thought is being given to issues related to the atmospheric dynamical core.
The thread on RC does not reflect much receptivity to the possibility of such ideas. It is to Gavin’s credit that he read the paper and spent some time at Air Vent. On the RC thread, he then proceeded to dismiss the paper, largely based on the reviews of their previous paper.
Pretty much the same way it does in any scientific discipline Judith.
A new theory is not accepted until it has shown itself to be a better explanation than the current accepted theory – i.e., scientific scepticism.
I’d say Gavin’s response is entirely in keeping with an appropriately sceptical stance.
Well of course. But that is different from dismissing something without a fair hearing and not allowing the paper in the published literature. The litmus test of a better explanation would require years and many additional studies, it doesn’t apply to a single paper.
And just because something is new, that doesn’t mean it should get a free pass.
I sometimes get the impression here that scepticism only applies to the more established science and credulity rules for all else.
Who said this is getting a free pass? It is a paper that is submitted for a journal. The authors are eager for paper to be discussed by the meteorological and climate community. I am providing a forum for such a discussion (RC obviously is not).
And I think Gavin participated in that discussion.
That it doesn’t get it’s own thread at RC isn’t a cause for complaint.
Your judgement seems to fail you when it comes to things RC.
Lets stop talking about Gavin and RC, but focus on the paper and the broader topic of how the community responds to outsiders and new ideas.
This is a great opportunity to have a meaningful exchange with the authors: two of them have shown up here.
As my first two comments did not pass the moderation, I repeat them here:
My colleagues and I are very grateful to Dr. Curry for her genuine interest in the science we propose and for the many critical and specific comments she made on our work. I could write a very long post what I personally think of Judy and her role, “a portrait of Judith Curry by an outsider”, but at this first large-scale introduction of our work at Climate Etc. I should concentrate on science.
In our work we derive pressure gradients associated with removal of vapor from the atmosphere through condensation/precipitation. Nothing except basic physical considerations are used. On the basis of our derivations, we show that on the scale of Hadley cell, the mass removal process produces just the pressure difference that is observed (10 hPa). The same process also quantitatively describes hurricanes and tornadoes, as shown in greater details in our previous works.
Judy refers to these estimates as “back-on-the-envelope”. This raises an interesting methodological question. In physics, you cannot solve a problem without such basic estimates that give you the overall magnitude of the effect you want to study. Once such an estimate is obtained, it signals you whether to study the effect further or not. For example, had we concluded that the pressure difference would be 0.01 hPa over 2000 km, we would not be interested in talking about this further. At present we do not have any doubts that the effect is as significant as we propose.
In this sense the arguments of the Editors of our previous ACPD submission were logically correct. They attempted to dismiss our work at the same level of basic physical arguments. Dr. Rosenfeld provided a basic (Judy’s back-on-the-envelope) calculation proving that condensation cannot reduce air pressure. The Editors agreed, and the paper was rejected.
OK, now we show that condensation does lead to pressure drop. What is the next basic objection? No objections at this level any longer, as I can see. The argument has shifted to “the models that do account for precipitation, show that the effect is small”. In other words, the models are complex and account for everything, reviewed by thousands, and we trust them, not you.
However, science is not a matter of trust. Our physics is so transparent that for people who have built GCMs and described the climate system in all its integrity, it should be easily possible to come forward and say: hey, Makarieva et al., where you say it is 10 hPa, it is in fact 0.01 hPa, because in your derivations you did … instead of …. If you did … which is correct, you would obtain 0.01 hPa.
In other words, one can summarize how the effect is included into GCMs at the “back-on-the-envelope” level and contrast this derivation with ours. That would be alternative basic physics approach and people could see which of the two is wrong.
Anastassia, apologies, new posters register, after registration, the comments should pass right through. For some reason, a number of comments got held up this morning, all of which i have approved. I will keep on top of this today to make sure that comments don’t get delayed.
Re: Tom Curtis October 23, 2010 at 10:40 pm
I think this is a very good example showing that many people do not understand what our work is about when trying to judge about the magnitude of its implications. Dr. Schmidt’s words about 0.6% of total pressure refer to the weight of vapor molecules in an atmospheric column. This is a store. The effect of condensation on pressure gradient that we are describing depends on flux — the rate of precipitation. The faster the vapor is removed from the atmosphere, the greater the resulting pressure gradient will be. Fluxes and stores should not be confused, as these are variables of different dimensions.
I would also like to encourage people, before spreading the word about 0.6% of total pressure as a negligible magnitude, to give a thought to the relevance of this magnitude in the atmospheric context. For example, Hadley cells (the major feature of Earth’s atmospheric circulation) are driven by a pressure difference that is of the order of 10 hPa, i.e. a meager 1% of total air pressure.
Dr. Curry evidences exceptional empathy to outsiders, laymen and innovators balanced with incisive power to nonjudgmentally relate new observations to interesting apt questions of hard science. (A)
I usually lack this knack, and often hamhandedly take the adversarial approach toward an idea in the hope that the patient promoter of the novel idea I as yet poorly grasp, and other commentators of goodwill bear with me long enough that what they’re trying to tell me sinks through my very thick skull, or we’ve trimmed enough rough edges off the original idea that it flies truer or is whittled away to something new, or nothing at all. (B)
It’s my observation that between our two extremes are many nuanced levels of people who — knowing more than I — have more to unlearn and possibly more vested to lose or reallocate and thus can be both adversarial and able to move the discourse in a direction in line with their own agenda while preserving the value of the old and new to some degree. (C)
These seem to have their own opposite completing the corners of the square in those who put the agenda above the science, and would rather collapse the discourse and lose both their own investment and the value of a conflicting new proposition than risk whatever new outcome. (D)
And all four in some way must be valid approaches. (A) is always a pleasure to learn from and expands participation in a field; (B) can test ideas for robustness; (C) often more quickly moves the discourse to a higher level; (D) can, for example, prevent small children from mixing common household chemicals in an Easybake oven and serving them to their siblings.
All four, just as much, may be dysfunctional in a community. As suggestions, (A) could be abused into endless digression either inadvertently or by unscrupulous manipulators; (B) is inflammatory and tends to a zero sum game; (C) can slip into any of the other three’s faults very quickly, and (D) leads to the lose-lose game almost inevitably.
I’m not sure how stable this A-D model of community receptivity is; possibly a transition network and some recursive or iterative method could allow us to simulate exchanges and outcomes involving outsider inputs and community reactions as a state machine.
But then, what would the utility of such a dynamic A-D model be?
Can sufficient skill be demonstrated for such a model to make support its study?
Could it forecast outcomes of questions, allowing an automated answer to be generated?
Would it predict adverse states and allow a moderator to redirect the exchange to a more desirable end?
Do most communities already have an intuitive model that exceeds the performance of this model and renders models of this sort moot?
Is it the structure of models, or the expectations of (type B/D) laymen unprepared to accept the limitations and parameters within which models operate, or both, that we ought seek improve?
Bart, interesting analysis. Individual scientists spread across A-D is fine and natural. It is a problem when a community with gatekeeping privileges (e.g. editors of journals, etc) tends too much towards dismissing new ideas. I am trying (sometimes singlehandedly it seems) to counter the overemphasis on dismissing new ideas and/or marginalizing outsider/newcomers that I see in the community caused by the political relevance of the subject (as evidenced by the fate of Hansen’s 2000 paper).
Prof. Curry,
Although you may feel lonely in your quest to convince your peers and the community-at-large to become more receptive to new ideas, you should not become discouraged. You mentioned in your last post:
” … Blog traffic has been pretty much optimal: sufficient to be rewarding (passed the 100,000 hit mark and approaching 5,000 comments) but not overwhelming.”
I don’t know how you might measure your progress towards your objective but, FWIW, the “hit” measure may include a number of your peers, as well as, others in the community. Only time may tell whether or not your efforts have influenced their thinking.
Well TTTM tried to raise a valid question, in a way so crosspatch that nobody managed to respond and it’s still not clear if TTTM understands the question.
“If you have a choice between a hypothetical situation and a real one, choose the real one.” — Joan Baez
Here’s a real one, copied over from the JC thread that references the JeffID thread where Gavin has participated, I have no idea where the conversation is actually happening now. But here’s a real bit of information relevant to modeling:
> Makarieava:
“… I think this is a very good example showing that many people do not understand what our work is about when trying to judge about the magnitude …. 0.6% of total pressure refer to the weight of vapor molecules in an atmospheric column. This is a store. The effect of condensation on pressure gradient that we are describing depends on flux — the rate of precipitation. …. give a thought to the relevance of this magnitude in the atmospheric context. For example, Hadley cells (the major feature of Earth’s atmospheric circulation) are driven by a pressure difference that is of the order of 10 hPa, i.e. a meager 1% of total air pressure.”
Now one percent across a volume the size of a Hadley cell is a lot more than something small subtracted from six tenths of a percent across a small volume the size of a weather system.
But can that draw rainfall into an area of intact forest, versus a patchwork or clearcut?
Wouldn’t surprise me.
Anecdotally, my favorite hang gliding mountain was densely forested, from half a century of overprotection by Smoky Bear, when I was learning to fly. It had huge big thermals, lovely cumulus clouds, and fairly consistent late summer afternoon rains.
Then it burned, along with a huge area.
Now, after a couple of decades of recovery, it still produces more intense smaller thermals, bumpier air, and only rarely develops its own afternoon rainfall.
There are pictures of Australia showing much the same kind of difference between rain clouds over farmland and forestland.
Changes in ocean plankton cloud nuclei could produce the same sort of variation.
Specifics, not hypotheticals, focus discussion. TTTM threw a wooden shoe into the gears by getting everyone distracted from a possibly fascinating idea.
I think I can focus the discussion on the part of the paper the authors consider unique.
Eq 37 is derived independently of any latent heat. Latent heat is a larger factor than any other in condensation, however heat conducts away easily.
The hadley cell example at the end of section 4.1, basic observed pressures and velocities are entered and a horizontal gradient is determined. The authors claim that this value is entirely independent of latent heat of condensation.
If you agree — the paper is validated and models need serious rework.
If not — the paper has a problem and models are reasonable.
Eq 37 is definitely the punch line for the paper. The key issue is the w/u; this implicitly assumes a lot about how the atmosphere responds to the mass disturbance. This isn’t something that can be sorted out out with paper and pencil. My sense of this is that the local pressure/density perturbations induced in the atmosphere would see rapid adjustments between the mass (pressure) and momentum fields by sound and gravity waves (which may have some consequence on the local evolution of the cloud, but not on any large scale circulation systems). The net effect of such mass disturbances on a scale that is of relevance to weather or climate is likely to be miniscule. On the other hand, large-scale column loss/gain of mass should have a noticeable impact on circulations; how noticeable depends on the magnitude of surface evaporation and precipitation and the horizontal scale of the mass perturbation. If we are talking about the scale of a single thunderstorm, then you will see quick adjustment of not much consequence. On the scale of a hurricane, I would think you should notice this. For climatologically large scale features like the monsoon and hadley cell, i think the effect should be noticeable, but I am having a hard time reasoning my way through to a place where this is a dominant effect. I think it could be an important effect for the haley cell and large scale monsoon circulations.
These are all very relevant points. We have a reply to these questions which I hope to present reasonably soon in a thread Judy has invited me to host. In brief, the effect is dominant in two cases: (1) when the length scale is very large (like Hadley and monsoons) and (2) when precipitation is very intense (like hurricanes and tornadoes).
For small scale slowly rotating eddies, of the scale of a single cloud, for example, Eq. 34 is not valid. This is implicitly clear from the assumptions made while deriving it: Eq. 34 is derived assuming hydrostatic adjustment, i.e. that the mass perturbation caused by condensation within the column has relaxed along the atmospheric height (which is a scale of a few km).
By the way, this responds to the concern of Dr. Schmidt that a single cloud does not immediately contract. According to our results, it should not.
“This is implicitly clear from the assumptions made while deriving it”
Anastassia, where are these assumptions stated?
The problem I’m having with 37 is that it is fed observed factors and the observed result is found. Doesn’t latent heat affect h gamma and might it affect u/w as well? The dimension of the storm makes a lot of sense but once the observations are plugged in, it could just be an accounting of hidden factors rather than just condensation. It’s all a bit confusing.
In Qiu C.-J., J.-W. Bao, and Q. Xu, 1993, a paper sent to me by email, they claim to include latent heat energy and condensation in the hurricane and get reasonable velocities. I’m not the hurricane expert though so perhaps you can explain it better.
Also, there was a plot provided which showed that taking the latent heat only model (which had reasonable velocities) and including the mass conservation from condensation, they were able to demonstrate the effect was minor.
Now the way they handled the mass loss is opaque to me from the way the paper is written. I’ve never read a hurricane model’s code but I wonder if they have done the pressure updates correctly. If there is any light anyone could shed on this, I would appreciate it.
I agree that there are many factors that are not included in 37. I don’t think this idea can be evaluated without using numerical models, because there are too many scales of motions and too many processes to consider. That said, existing numerical models won’t work for this since they don’t include this mechanisms and other compensating mechanisms are in play. So we will see what AM’s scale analysis has to say. I am also thinking of something that might be done with the pressure tendency equation to sort this out.
M10 see this as a paradigm shift in understanding due to 37. Ignore latent heat and use only condensation. The paper doesn’t teach the size of cell the eq applies to from my interpretation but I’m not convinced that the hadley cell results are independent of latent heat.
Still, whether the interpretation is right or not, the result they got is correct and that should say a lot about how we interpret weather. I do FEA for optics and have done mechanical FEA (both from scratch) it is very simple to include effects in these things without fully understanding the implication. I’ve designed multiple lenses without having any idea what their shape would be until the process was done. This work (if properly interpreted) gives insight into why it works, and that is a definite improvement.
That is why I keep going back to the potential non-independence from latent heat of the hadley cell example.
I agree that a much more extensive analysis is needed. i regard this paper as provocative, the “first word”, certainly not the “last word.” But that is how science works, and I also appreciate the fact that this idea and paper came from outside of the mainstream climate community.
Over at the Air Vent Anastassia writes (reply #151):
“My point in #54 was to draw attention to the glaring mismatch between the level of performance of the theory and models. My reading is that in 1980 Held and Hou offered what could have become THE solution to circulation problem. They satisfactorily reproduced the length scale of the Hadley Cell starting from a few fundamental assumptions and basing, conventionally, on consideration of differential heating. However, the circulation intensity that the theory yielded proved to be one order of magnitude lower than in reality (see page 662 in Schneider 2006). As far as I understand, had the velocities coincided in that paper, one could celebrate solution of the general circulation problem. But they did not, and the solution, to a varying degree, remains elusive until now. To us it is not surprising, as we state that an essential process has been overlooked.
In the meantime, I feel that already in the 1980s (and much earlier) GCMs were reported to be good enough to reproduce the velocity field. I am not aware of any published concerns that the models showed a one order of magnitude weaker circulation. From these observations one is led to conclude that GCMs apparently did not need the theory, their level of performance was reached by other means. At least it is my reading of the situation.”
This leaves me extremely puzzled. It seems that the theory on which the models are based is known to be an order of magnitude off, still the models do not show this weaker circulation. How is this possible? Anyone care to explain what these “other means” are?
This point about the Held and Hou (1980) model that Anastassia makes is well-taken, but the deficiency in the simple physical approach should not lead one to conclude that GCMs in the 1980s were in capable at simulating the large-scale atmospheric circulation. Instead, the ‘take-home’ message should be that there are other processes at play, besides differential latitudinal heating that may be ‘driving’ the Hadley circulation. What has traditionally been invoked to explain the weakness in the Held and Hou model is that their simple model lacked latent heat release, which had it been included, should lead to an intensification of the Hadley circulation.
James, thank you for clarification. My reading of the situation is that Held and Hou correctly estimated the efficiency of the heat engine caused by differential heating and found it to be low. This efficiency depends on turbulent heat conductivity. Latent heat flux is a fraction of total solar power in the order of several dozens per cent. So, accounting for latent heat could have increased total differential heating gradient by no more than twice (all latent heat captured by evaporation in the tropics is released at the equator). Provided that turbulent heat conductivity is the same, this could not raise the intensity of Hadley cell by the needed one order of magnitude. In fact, I consider that paper a proof that differential heating is not a major player here.
I would also like to point out once again that it is not plausible to compare fluxes of heat (associated with latent heat release) to fluxes of work associated with mass removal. For the release of heat we have dQ = cvdT + pdV, where the term cvdT under atmospheric conditions has the meaning of turbulent heat conductivity (over horizontal plane). So dQ can be very large, but so is the heat conductivity, leaving work pdV negligible. Hence the theoretical problem of quantifying the efficiency of the atmosphere.
Conversely, pressure change due to mass removal of ideal gas is pure potential energy, Vdp = -pdV (at RdT = 0). It is the unique property of ideal gas, which is derived from its equation of state, not the 1st law.
Thank you for your comment, Anastassia . First, I want to say, I’ve really enjoyed this paper and others by your group. My concern with your latest paper is that I think you may be overstating the impact that this process has in driving large-scale atmospheric circulation. I’ve gone through your analysis and find no errors in the mathematics and/or physical interpretation using an idealized atmosphere. However, as you bring up in Section 5 of your paper, the real atmosphere does contain structures (both large and small) where the magnitude and scale of the winds are nontrivial and that don’t appear to be appropriately described by this mechanism. You specifically mention land/sea breezes and dust devils and say they are trivial compared to the general atmospheric circulation, but what about the stratosphere or higher atmosphere(where condensation processes are negligible), but wind speeds can reach very high values (> 50 m/s). These phenomena and others appear to be adequately by the current differential temperature paradigm. So in my opinion, the current disconnect that exists in this latest paper is in translating this mechanism to the real atmosphere. I think to fully understand how important this process in the real atmosphere, it has to be done in a modeling study using the primitive equations where condensation-induced pressure gradient is specifically included. Until that’s done, I think it is inappropriate to conclude that this mechanism can adequately explain many types of atmospheric circulations where condensation/evaporation occurs.
James, thank you for this comment. Let me tell you that all your previous comments and thoughts on the topic have been very much appreciated. Regarding the problem of overstating the effect, I have to say the following.
Every paper carries some new basic stuff and some conclusions/perspectives provided by the authors. The authors have spent more time than anybody else thinking on their basic stuff (otherwise it would not be new) and its relations to other phenomena. Not all of this comprehensive knowledge can easily be put into a single paper. But I think that the authors (not ourselves, but generally, the authors) must be allowed to give their perspective, even if it is a far-reaching one. It can stimulate other researchers. After all, the authors could die tomorrow, and if later work would show that they were right, it would be sad to lose the original perspective.
Let me give you one example. In our 2007 paper we gave a couple of sentences’ mention to the fact that the pressure difference caused by vapor condensation should spread onto the horizontal dimension to produce realistic pressure gradients (HESS, p. 1023, third paragraph). At that time we did not present any proof for that statement (and we did not have it). But had our 2007 paper received a more constructive interest in the meteorological community than it did, we are certain that the results we are now presenting could have been obtained by someone else, perhaps even earlier and in a more complete and elegant form.
The problem is that the current worship of peer-reviewed publications in climate literature does not actually allow for a debate, for daring a perspective, for exchanging opinions, everything that comes under the peer-reviewed label is confused for some “ultimate truth”. Informed scientific opinions, instead of being considered as incentives and suggestions to think in one direction or another, are confused for political statements.
You say that “until that’s done, I think it is inappropriate to conclude that this mechanism can adequately explain many types of atmospheric circulations where condensation/evaporation occurs.” We are convinced, from all we know, that this mechanism can adequately explain things. This said, we fully respect your opinion. As everybody else, we can be mistaken. What we say in the paper is that the effect is important, it should be studied rather than ignored or dismissed as it has been until now, and that one cannot claim a fair understanding of atmospheric processes without taking this effect into account.
In practical terms, this means that for atmospheric scientists it should not be possible to continue theoretical research as if nothing happened. I fully agree with you that further research is needed, in fact, it is what we are calling for. But one cannot say — explain everything to us, then we will look at your stuff seriously. One cannot expect that a few authors will solve ALL atmospheric problems. If it were possible, it would be unclear if the huge funding globally allocated to atmospheric science is at all appropriate.
“But one cannot say — explain everything to us, then we will look at your stuff seriously.” This is THE stumbling point today, not just in climate science, but everywhere those in authority (and the many lay “defenders of science”) believe in the consensus rather than the very real possibility of new knowledge, that would overthrow that consensus, by invalidating a crucial part of it. In my own research into the objective origin of the “ancient mysteries”, which it turned out directly affects one’s understanding of all the earth and life sciences, I recognized, at one point early on, my own looking in slightly the wrong direction, and in writing about that short episode later (ten years ago now), I compared it with the old joke about the man who looked for his lost coin under the bright light of a street lamp, instead of across the street in the dark, where he dropped it. I see this “looking in the wrong direction” (indeed, insisting upon doing so, even in the face of positive knowledge being brought forward against the consensus) time and again, across many fields of modern science, but especially in the earth and life sciences. “We don’t want to look there; come over here in the light (of what we think we know–what we want to believe).” Even so has the vaunted self-correction of science been long stymied, across the board. Truth is where you find it, not where the “experts” dictate it must be.
James, turning to science, let me give my view on winds exceeding 50 m/s in the higher atmosphere.
As Jim D pointed out here, it is not possible to conceive a circulation consisting of converging ascending air only. In our work we quantify the power of the process that drives the circulation by creating pressure gradient in the lower atmosphere. As an air flow is initiated near the surface, the distribution of momentum flux leads to appearance of a closed circulation, there is no other opportunity. Vertical and horizontal re-distribution of momentum fluxes leads to the appearance of horizontal pressure gradients in the upper atmosphere that will have an opposite sign compared to what we observe at the surface.
(Forgive me this everyday example: if you have a vacuum-cleaner, it sucks the air in in one direction only. This is the circulation driver. However, the resulting circulation of air will include branches where the air moves both along the direction the driver works along as well as against it.)
Therefore, condensation in the lower few kilometers will make the air move above the condensation level as well. Since apparently the two horizontal parts of the circulation (the lower wet and the upper dry) should be characterized by approximately the same power (having one and the same physical driver), one can expect pressure gradients in the upper atmosphere to conform to the same constraints as in the lower atmosphere. However, in the lower atmosphere there is surface friction, while in the upper atmosphere winds are close to geostrophic (and, hence, higher than at the surface). If you take the pressure gradient that we obtained for Hadley cell and calculate geostrophic wind velocities, with an account of the fact that air density is several times lower than at the surface, you will easily obtain winds in excess of 50 m/s. (See Eq. 39 in the paper, which gives a lower estimate).
These will be “dry winds” driven by mass removal of vapor that takes place in the lower atmosphere.
I have great difficulty visualizing the circulation postulated with w and u in section 4.1. If there is a pressure gradient into a saturated column at all levels, the circulation is not sustainable or steady. It has to be outward directed at least at the top.
Aside from this, the model suffers some of the same drawbacks as older GCMs by not allowing suspended condensed water to contribute to the hydrostatic pressure. This loading effect has been shown to be important in reducing hurricane intensity. Newer GCMs that carry condensates are better in this respect than the older ones that rained everything to the surface immediately, as I think this model is doing.
Not quite sure i understand your 2nd para. GCMs with microphysical modules that carry prognostic equations for condensed water do a much better job of simulating precipitation. But this condensed water does not contribute to the hydrostatic pressure in most models that I’m aware of?
You may be right, as there are technical issues that favor using the dry pressure for the coordinate, while integrating a separate full pressure including moist species and vapor for the pressure gradients, and not many models correctly make this separation.
Question?
How can you have theories on the atmosphere when no one has figured out how it was created and how it works exactly?
It is an extremely complex system that generates pressure through rotation. Gases regenerate from this planet compensating for the bleed off through friction and space travel.
We need to learn to walk before trying to fly!
If not, every theory generated will be incorrect as something was left out.
I looked at the parallel Air Vent thread on this, and will give my take on the pressure-drop issue. You have to separate two pressures here, the hydrostatic pressure, which you get from integrating the mass above, and the “full” pressure which is the one you get from the local gas law. These are not necessarily equal in the real atmosphere. Hydrostatic models only have the former, while the latter responds immediately to local physical processes such as heating and phase changes.
In the real atmosphere, condensation immediately reduces the full pressure due to loss of gas pressure from some vapor that is condensed. Hydrostatic pressure is not affected because the mass is still there in liquid form. This change in full pressure leads to a quick response from neighboring areas to fill the void (effectively sound wave speed), thereby increasing the density until the pressure is equalized. The final result is that the volume would have higher density IF it hadn’t also heated by latent heating at the same time. The heating has an opposite and greater effect on the pressure than the vapor loss. The net effect is a higher pressure, followed by mass divergence by sound waves, and finally reduced density as expected for the warmer air.
Hydrostatic models short-cut the sound-wave part by just assuming the new density immediately is a function of the new temperature and vapor content, effectively equivalent to having infinitely fast sound waves providing the necessary mass transfer to change the density. They do end up with the correct air density which is a combination of higher temperature and less vapor that contribute oppositely to it.
Thus I see the pressure drop due to vapor loss as part of a tandem with the pressure rise due to latent heating, both of which are accounted for if the hydrostatic model air density includes vapor content in its calculation.
The pressure drop effect, while present, is a very temporary situation that finally leads to the new density that hydrostatic models respond to. I see it as part of the hydrostatic adjustment process described more fully by Bannon.
Agreed. The important impact is when mass is actually loss from the column in the form of precipitation.
Is precipitation drag included?
Hydrostatic models can include precipitation drag (effectively the same as weight of condensate) as a part of the hydrostatic pressure, but since they don’t have the full vertical motion equation, its feedback to the dynamics is limited to that.
Jim D:
Thank you.
Your excellent summary helped my understanding of the paper especially the issues with eg. equation 37.
Richard
OK, I’ll try to make a brief point continuing Tim the Tool man’s line…IIRC, Willie Soon has asked the same questions and made the same points before.
If one component in a climate model is ‘wrong’, does it invalidate the whole model or only that component?
Shub, it depends on which ‘component’ is invalidated. If the physics of gases and the radiative transfer calculations are found to be “wrong” then there’s a whole heap of stuff to be done or redone. On the other hand there are many, many observations, data and modelling components which would only require new or extra numbers to be factored in to current theories or calculations.
And then there’s the following issue. If the science and the models based on that science are wrong in some drastic, fundamental way – how do you come up with the theory needed to explain the observations. We have the data, we have the graphs, would we be left staring helplessly at them?
If the Arctic sea ice continues to melt and oceans continue to acidify and birds and beetles continue to move habitats – does anyone have an alternative theory ready to step up to the plate as the new paradigm for explaining all the records and observations we’ve so far accumulated?
Adelady:
In hope of returning to consideration of the possibly important paper by
Makarieva et al. I provide an answer to your question. You ask:
“If the Arctic sea ice continues to melt and oceans continue to acidify and birds and beetles continue to move habitats – does anyone have an alternative theory ready to step up to the plate as the new paradigm for explaining all the records and observations we’ve so far accumulated?”
Yes, natural variability.
All those effects have occured prior to anthropogenic emissions of GHGs so could have – probably do have – the same causes as they always did.
At issue is gaining a greater understanding of the climate system and the mechanisms responsible for its variation. That understanding has value whether or not AGW exists to a real, discernible or potentially catastrophic degree.
And consideration of the paper by Makarieva et al. has potential to contribute to that understanding whether that consideration induces acceptance or rejection of the arguments by Makarieva et al..
Conversely, rejection of the possibility that our present understanding(s) may include major errors is a rejection of the fundamental principle of science that new information and/or interpretations can overthrow existing understandings (i.e. knowledge). And this principle is why the response of ‘gavin’ to the question from ‘TimTheToolMan’ induced outrage from several (including me).
Richard
Natural variability? That would be fine – but at the moment we have a problem with that.
“Normal” natural climate and its variability is currently explained using the radiative-transfer-physics-of-gases-etc as the underlying explanation for why the earth is sometimes covered in ice and sometimes suitable for agriculture and human society. That same concept (and all its friends like Milankovitch cycles and the rest) underpins the calculations verified by current observations and used to project likely future conditions.
Personally, I am really, deeply anxious for someone to come up with an alternative construct that both describes and explains current and past climate and gives us confidence in a better future than the one we seem to be facing. Natural variability only works as an alternative if it’s the outcome of another (as yet unknown) mechanism that doesn’t give the warming result from a large pulse of greenhouse gases into the atmosphere.
The ‘overthrow’ or ‘house of cards’ approach is very, very attractive, but my view of really complex science like climate or human biology is that a root and branch revolution is pretty unlikely. No matter what medical scientists tell us in the future, we know a fair bit about how we grow and develop and how various body functions go wrong. Even if the rather new idea that obesity is a metabolic safeguarding mechanism against the toxic effects of too much fat in the diet turns out to be right (or wrong) – it doesn’t alter the view that obesity is dangerous and should be avoided because it will cause serious problems for the heart and other organs.
I’m happy with ‘it’s natural’ as long as the ‘natural’ has a different physics explanation from the current one.
Shub, it depends. Other degrees of freedom may pick up the slack, or not. No simple answer in a complex model.
Dr Curry
I’ll try to elaborate my viewpoint and perhaps reframe my question, since similar concerns arise in the area of my research as well.
Let us say a subcomponent of a model is part of the so-called dynamical core, and there is something wrong with it. Does it invalidate the model?
My argument would be: it does.
After all, a ‘model’ is two things – the integrated specifications of individual subcomponents, and the belief that the components interact meaningfully to produce valid output.
A breakdown in one subcomponent affects that component, and by its interaction with others, affects their inputs/outputs, and therefore affects overall model outputs as well.
Saying at this juncture that subcomponent breakdowns do not matter, if and because overall output remains unaffected, is atleast in my view scientifically unacceptable because it violates model integrity and validity.
We must not only know the ‘right answer’, but also for the right reasons as well.
adelady,
Climate models at present are unable to replicate seasonal variations. Why?
I agree that the dynamical core is the most important part of the climate models. All sorts of small approximations have been made, and compromises are made to make the solution of the equations tractable on the computer. So the issue is to what extent whatever is ‘wrong” with the model makes a difference in the simulations in terms of the task for which the model is designed. This can only be evaluated objectively by fixing the model and then comparing it with the original model version.
Thanks Dr C and Nick,
So the question now is, and I am aware this might be early to be asked,…does the fact that the mechanics Dr AM describes, are not included in present GCM – as she has claimed – invalidate them?
I am sticking to the tack that models should get the right answer for the right reason – ’emulation of physics’, and, let us say, do not place any premium on models getting outputs matching reality and therefore being declared valid.
I am of this opinion too. Afterall in theory a neural network could be trained to model our weather quite nicely given all the data we have. There would be no “science” or physics involved at all.
I wouldn’t like its chances of getting any reasonable results from parameters that go outside its training, though.
I see the GCMs in this way too. In a general sense, the right answer for the wrong reasons is going to be a wrong answer under other conditions.
This is a key point. In our case it is very easy to give an example. GCMs do not predict practically any signficant consequence for the water cycle from deforestation even in such regions as the Amazon. Regional water problems are broadly ascribed, as everything else, to climate changes due to CO2 accumulation. To stop CO2 accumulation, one is called to grow biofuel or reduce emissions in some other way. In contrast, our analysis suggests that the way to sustain or restore the regional water security goes through preservation of natural vegetation cover, while industrial carbon emissions are irrelevant. We thus have drastically different recommendations as to what to do depending on how we understand the physics of atmospheric motion and the role of condensation/precipitation in atmospheric dynamics. This is but one example.
Another example is that without understanding the driving force behind atmospheric circulation it is not possible to estimate thermal sensitivity of the climate regime, as the latter depends much on latitudinal transport of mass and energy.
AM “while industrial carbon emissions are irrelevant.”
I read the abstract only. I have a personal view that maintaining and restoring forests is essential for a whole host of reasons, climate among them, so I’d agree with you despite my lack of any scientific capacity to review that paper.
However, I’m at a bit of a loss about the quote above. Do you mean just that forestry, and the associated water cycles, can be regenerated without attention to CO2? Or are you saying that forestry alone can sequester enough carbon to counteract ocean acidification as well?
Adelady, thank you for posing these questions. In brief, in our work we show that relying on the regulatory potential of natural ecosystems it should be possible to achieve environmental stability without the proposed reductions in CO2 emissions. You can read our website and perhaps this paper in particular to have an idea of our standings. Perhaps discussing them in greater detail here would take us too far off-topic.
TimTheToolMan:
The problem with neural networks, children and dogs, is that you know what you have taught them, but you don’t know what they have learned.
Alex
Shub,
I’m not expert on GCM’s, but I do know something of CFD. Whether something goes in the “dynamic core” or the outer region is not really an issue of accuracy, but stability. As long as it is updated somewhere in the timestep, and you don’t get spurious oscillations, it’s OK.
Of course if it’s not in either place (and that I don’t know), there’s a problem.
Robert H. Essenhigh has formulated thermodynamic models that may be of interest in comparing these hydrostatic models:
“Prediction of the Standard Atmosphere Profiles of Temperature, Pressure, and Density with Height for the Lower Atmosphere by Solution of the (S−S) Integral Equations of Transfer and Evaluation of the Potential for Profile Perturbation by Combustion Emissions”
Energy Fuels, 2006, 20 (3), 1057-1067 • DOI: 10.1021/ef050276y
I find very few citations to what appears to be a substantial effort on improving the thermodynamic atmospheric model.
David, i can’t get easy access to the paper, can you email a copy? i would be interested in looking at this. thx
Sent.
Essenhigh’s formulation is applied by Sreekanth Kolan in his 2009 thesis:
Study of Energy Balance between Lower and Upper Atmosphere
Hi all,
I am working through the paper as best as I can and I have not got very far but a couple of things are bothering me.
The first has do do with timescales.
As I understand it, most thermodynamic processes occur very quickly; so that the timescales of practical heat engines are longer than the underlying processes, or at least the are treated as such upto the level that I have seen illustrated. I am really not sure that is the case where condensation and evaporation are concerned.
If one cannot normally expect condensation to readily occur then the left hand side of equation 11 (after substitutio of eq 13) could be read as the potential for a certain amount of condensation/evaporation to occur; not the guarantee of it occuring without stating a timescale. As I read it, the equation supports, amongst other things, that devices such as cloud chambers do work as described. They have the potential to produce condensation after adiabatic expansion.
The second has to do with inertia.
This is more tricky, but at its heart is a suspicion that the possibility for a parcel to become supersaturated (in amongst a large volume also becoming supersaturated) and then commence to rapidly condense (at the same time as does much of the surrounding volume) leading to an increase in pressure due to its inability to move adjacent volumes out of the way quickly enough. That is its, ability to do work at a sufficient rate is hampered by the inertia of the masses that surround it.
Comining the two into an atmosphere scale cloud chamber one might have adiabatic expansion with declining pressure but no condensation, then rapid condensation accompanied by a rise in pressure, followed by adiabatic expansion lowering the pressure. The final pressure being less than the initial pressure. In that case the outcome would be as the paper suggests but the process somewhat more convoluted.
I simply do not know the timescales on which these effects play out, nor if a have a sure grip on the topic in general.
If such things are issues then perhaps rates of change need to be treated explicitly.
Anyway that is as far as I have got. Onwards and upwards.
Alex
I should have added that I am not suggesting invalidation. The power for condensation to do signifiacnt work is not being challenged in any way, possibly the converse. I am just seeking clarification on the timing of events, i.e. on process not outcome.
Alex
I think the arguments are made in the context of steady state. I don’t think their arguments are useful for cloud scale processes, but may appear in larger (near) steady state circulations like monsoons.
This is a comparison of the approaches to deriving Eq. 34 (or its analogue) in our paper and in Curry and Webster (1999) [CW99]. This is my first attempt on this, and I very much would be interested in your comments. My apologies to Judy in advance if I have misinterpreted something in your text.
In Eq. (5.28) [CW99] one calculates the change
, where 
is vapor pressure, 
is saturated vapor pressure, as an outcome of two processes: ascent of water vapor at vertical velocity 
and condensation of vapor. The physical meaning of the first term, 
, is to describe the magnitude by which vapor pressure would have increased in the absence of condensation, while the second term, proportional to 
, describes how much of this excessive moisture has condensed. In the stationary state the terms sum up to zero, so any of them can be used to gain information about the condensation mass sink. (Note that 
here has a different meaning than in M10)
In the stationary case that we are considering
(vapor is saturated). Multiplying both parts of equation (5.30) by 
and noting that 
in the stationary case, we obtain the following expression for the local change of pressure as determined by local concentration change:
Now the important point is that here
is vapor pressure gradient which would have taken place in the absence of condensation. As we state in the paper (p. 24033, line 1), such a term is necessary to discriminate between pressure changes due to condensation and pressure changes due to adiabatic expansion. So 
is such a reference term. CW99 presume that in this case water vapor would have been in hydrostatic equilibrium, Eq. 5.31. However, this assumption does not conform to the observation that the non-condensable gases of the dry air do not each conform to hydrostatic equilibrium. They are well-mixed, such that if the dry air mixture as a whole is in equilibrium, but each component of it is not. Otherwise the molar mass of dry air would change with height. Therefore, instead of Eq. (5.31), in M10 another equation is used, which supposes that in the absence of condensation water vapor would have been mixed as other dry air gases. This means that instead of Eq. 5.31 we have 
, where 
is total air pressure that conforms to hydrostatic equilibrium, and 
is the vapor’s share of the total local air pressure. Using this equation, the above equation becomes
which, taken with the minus sign, is Eq. 34 formulated in terms of pressure instead of molar density, see the right-hand part of Eq. 37 (note that
).
In the above post should be “one calculates the change
(
).
Sorry if I’m missing something, but if S = e/e_s and you assume e=e_s, doesn’t that make S identically 1 and dS/dt=0?
Would it be possible to write out Eq 5.30?
I would suggest that you asked Judy to email Chapter 5, I appreciate your difficulties, it is indeed not transparent. e stands for vapor pressure which would be observed if vapor locally behaved as non-condensable, es stands for saturated pressure. In the stationary state locally e=es, but de/dz is not equal to des/dz.
Eq. 5.30:![dS/dt = [e_s de/dt - e de_s/dt]/e_s^2](https://s0.wp.com/latex.php?latex=dS%2Fdt+%3D+%5Be_s+de%2Fdt+-+e+de_s%2Fdt%5D%2Fe_s%5E2&bg=ffffff&fg=333333&s=0&c=20201002)
Eq. 5.31:
Eq. 5.32:
Eq. 5.33:![de/dt = {-eg/(R_vT)]u_z](https://s0.wp.com/latex.php?latex=de%2Fdt+%3D+%7B-eg%2F%28R_vT%29%5Du_z&bg=ffffff&fg=333333&s=0&c=20201002)
Eq. 5.33:![de/dt = -[eg/(R_vT)]u_z](https://s0.wp.com/latex.php?latex=de%2Fdt+%3D+-%5Beg%2F%28R_vT%29%5Du_z&bg=ffffff&fg=333333&s=0&c=20201002)
A further comment on physics. The derivation presented in CW99 is genuine as it shows that the condensation rate should be an independent equation not derivable from the continuity equation. There is no information about Eq. 5.31 in the continuity equation. The physical logic of derivation in CW99 and M10 is the same — neither of the two equations can be obtained from the continuity equation, they employ an independent reference term. The difference is due to the fact that vapor molar mass is different from other gases, but this is not essential to the processes discussed in our paper, being rather a minor numerical correction. For atmosphere containing gases with equal molar masses, the derivations in CW99 and M10 coincide.
CO2 Residence Time
CO2 residence time is another critical factor driving the control/no control policy issue. See:
Potential dependence of global warming on the residence time (RT) in the atmosphere of anthropogenically sourced carbon dioxide
RH Essenhigh – Energy Fuels, 2009, 23 (5), pp 2773–2784
DOI: 10.1021/ef800581r
I recommend you add this order of magnitude difference in atmospheric CO2 residence time and the control/no control consequences as a major issue in your evaluations, with an uncertainty “Flag” to go with it.
This issue is pretty interesting, i don’t know much about it but would like to look into it at some point.
As several commentators raised this point, I would like to comment on the hydrostatic equilibrium/liquid water load issue and how it applies to our work and results. The physical content of the pressure gradient obtained in Eq. 37 is as follows: vapor has condensed in the rising air, but dry air has filled the shortage such that the air column remains in equilibrium, while the surface pressure decreases.
What sort of vertical pressure equilibrium are we talking about? Stationary shortage of vapor in the upper atmosphere due to condensation is of the order of
, i.e. of the order of 1% of total air pressure. If this shortage had remained non-compensated, one would witness huge vertical velocities w of the order of 
, possibly observed only in tornadoes. The routinely observed deviations from the hydrostatic equilibrium in the atmosphere are therefore much smaller than 
and do not impact the derivation of Eq. 37. To summarize: what we mean by the hydrostatic adjustment assumption in our work, is the statement that the disequilibrium pressure in the column is much less than 
.
I dwelt on the problem of the water load at the Air Vent here. The bottom line is that the weight of the liquid moisture that steadily remains in the ascending air (even in intense storms) is much, much lower than
, often by orders of magnitude. Therefore, our assumption of hydrostatic equilibrium AND the existence of a water load in the rising air are perfectly compatible. The water load and associated friction will certainly impact the resulting velocities.
One more point that I believe is important is that many people seem to think that until liquid water precipitates, the atmosphere remains in hydrostatic equilibrium. This is not so, of course, — had it been so, why should the droplets fall down or gas expand upwards at all, if condensation “per se” would not immediately disturb the equilibrium. Namely this break of equilibrium brings about the relaxation processes: the drops fall down, the air ascends.
This would be a great thread if people could keep their opinions about Gavin, RC, and the general state of climate science to themselves.
Really, please, you are not saying anything at all new or interesting. Keep it to yourselves or at least take it elsewhere so you don’t muck up otherwise exceptional threads.
Please show some respect for our host and the other serious contributors.
Yes, comments about Gavin and RC are now off bounds for this thread, although there were some early comments owing to my bringing up some comments at RC, which have since been deleted from the main post.
I’m sure that it’s ok to point out that Gavin has an interesting response to the latest AV thread here.
http://noconsensus.wordpress.com/2010/10/25/345435/#comment-39392
Jeff thanks for the update. we would appreciate a summary of what you have come up with on your thread.
While there is absolute merit in the examination of condensation and evaporation in cloud formation , and also the formation of ice /snow in the same clouds there is a factor that is being ignored. This is the energy absorbed by clouds long after the sun has gone below the visible horizon. As the earth is heated 24/7/365.25 on one portion or another, clouds being up to 30,000 feet above the surface are heated until about 10:00pm the maximum cooling of the atmosphere occurs about 1:00 am at which time the upper atmosphere again starts warming because of sun light bending by refraction from the east. This is based on IR radiation measurements over a period of several months. I’m sure that satellite data on upper atmospheric temperatures taken every time a satellite circles the earth will show this phenomena more accurately than my ground based reading.
My measurement to date include reading in all quadrants of the visible atmosphere and from both clouds , partially cloudy and cloudless sky. The readings are significantly different.
I am total bothered by the continuous use of the term “climate change and climate models” how the hell can a formula or a series of short term measurement project a condition that is made up by thousands of weather day end to end for one location. There are thousands of different “climates’ in this world and no one has proven that any of them have change over the last 150 years. There may be some temperatures increase but that has been going on for the last 10,000 years since the last ice age.
What is causing this warming? Changes in earth or solar orbits caused by changes in planetary alinement? Has the earths core been changing surface temperatures? There is no “creditable data showing that the “ghg effect exists”.
Dr. Curry:
You say, “Bottom line: it is challenging for an “outsider” to get a paper published that poses a major challenge to the status quo. Insiders are less likely to challenge the status quo, so outside challenges should be welcomed and considered carefully. ”
Not to derail a scientific discussion that I’m completely unqualified to comment upon, but you once again make a salient point about the bureaucratic inertia of science and the “not invented here” syndrome that infests all bureaucratic organizations.
This unwillingness on the part of some not to rock the boat can almost certainly be traced to the not-unnatural desire of those who have formulated their own theory for their theory to prevail. After all, and unfortunately, one’s continued existence as a member of the scientific community depends upon funding, and having one’s pet theory overturned by some “outsider” is not generally an upwards career path.
This is a sad but true characteristic of most human endeavour. Ego overtakes altruism, and innovation suffers. Nothing new, it’s been going on since Ptolemy was a boy. Unfortunately, with government control of the purse strings becoming more and more prevalent, science appears to be serving the political interests of the political elite more and more.
In the first section of the paper, I’m not sure the form of the 1st law is correct – the reference is from Gill, but without a page number it is hard to find a specific equation in that book (it’s a very “dense” text). The latent heat term is written as L*dgamma = L*d(p_v/p), which would imply a change in dQ from latent heat if only the pressure from the dry component of the air changed. That doesn’t make physical sense to me. I think that term should be (L/p) * d p_v, instead. I’m not sure what effect it would have on the rest of the section, but perhaps that would explain why the results conflict with Poschl’s comment on the earlier paper (and Gavin’s calculation in the other thread).
(Apologies if this has already been discussed and I overlooked it… the threads are a bit long)
Judith
You wrote that you were interested by my opinion on Anastasia’s paper.
I took a bit of time because I like to derive the equations myself so that one can after that only focus on hypothesis and not on technicalities.
Section 2 is classical thermodynamics. It reads like an exercice: “Consider a small condensation. Apply Clapeyron and find the sign of dp”. Nothing exotic involved, the results are clear for both the reversible (dS=0) and irreversible (dS>0) case.
Aside for Aronne: No text book necessary. For latent heat generated by a mass dm you have dQ= L.dm
Now as m=M.Pv/P, dm = M.d (Pv/P). As Anastasia uses molar quantities, it’s L.d(gamma) as in equation 1.
Section 3 is hydrostatic equilibrium. The maths are again straightforward and no exotic hypothesis involved. An estimation of dp for the hydrostatic case is done and connected to the Section 2. So the results obtained for the hydrostatic case seem again consistent and robust.
My comment is that it is only the hydrostatic case and this is a hypothesis I keep in mind. It is important because the atmosphere is not in hydrostatic equilibrium. Actually one could say that the whole atmospheric circulation is only taking place precisely because the atmosphere is not in hydrostatic equilibrium and scale considerations change nothing on that fact.
That means that I will look with high suspicion on every use of the Section 3 results on a system which is not covered by the Section 3, namely a system not in hydrostatic equilibrium like the atmosphere.
Section 4 is the real matter.
Eq32 and Eq33 is just classical mass conservation. Anastasia having chosen a 2D model, the good news is that 2D Navier Stokes (which will have to come in the next paper or the one after that) is tractable.
The bad news is that her velocity field is not divergence free so it will have to be compressible Navier Stokes.
As an aside, it is not very consistent to have an explicitely compressible flow and use on it incompressible Bernouilli (like f.ex in 4.3).
I was unable to derive by myself Eq 34. It is clearly not a result of what came before. I broadly understand what it means so I accepted it as a given.
The paper would certainly win much if there was a derivation of Eq 34.
Once 34 admitted, 35,36 and 37 follow easily.
Well as for 37, what follows easily is :
dp/dx = (dPv/dz – Pv/PdP/dz).w/u
But this is already enough to show a p gradient along x what was the purpose.
The jump to P.w/u .gamma/H(gamma) is using the hydrostatic equilibrium assumption so, as I said, I look at it with deep suspicion :)
In my opinion the paper finishes more or less here.
Surely the next step of Anastasia and Victor will probably be to cross from equilibrium considerations (thermodynamical, hydrostatic&Co) to dynamical cosniderations.
I can’t see how it could be done else than by Navier Stokes. Probably 2D and steady states first.
It can in my opinion also be only after this step that one could tefficiently talk about GCM. Of course I am sure that the GCM will have missed this issue like many other but this paper is probably not (yet) enough to prove it.
So only briefly to teh following sections.
I was very interested by 4.4 and 4.5 even if they are more qualitative and lack the mathematical rigor and accuracy that Anastasia showed in the the first parts of the paper.
Even if I don’t think that the evaporation is a surface (only) process because water evaporates within clouds too, I believe that there is indeed an asymetrical dynamical game between condensation and evaporation.
It reminded me the “conformational entropy” where we discussed the fact that the specific distributions of spatial states matter in spatio-temporal chaos.
Clearly averaging spatially evaporations, condensations and precipitations like what is currently done in climate “science” is behaving like an elephant in a china shop.
As conclusion I would say that this is a very well written paper.
There is rigor, good maths, no exotic assumptions, the results are correct and quantitative up to 4.1 (imho) and an enormous amount of goodwill to explain and communicate.
It would be pleasant if all papers were written like that.
thanks much tomas.
I was able here to get an equation with terms like those in Eq 34 by eliminating the velocity gradients from Eq 32 and 33. But the terms involving horizontal density gradients don’t go away, nor do I think they should. I think that if there are situations where those horizontal gradients can be ignored, they should be specified.
The situation is less clear because of the paper’s insistence that 34 does not follow from 32 and 33 (cons mass) but is independent. I think other physics would bring in other coefficients (eg specific heat) which are absent from 34.
There is also the matter of the definition of Nv. As stated later in the paper, 33 is a normal conservation for water vapor species. But Nv is defined as the molar density of saturated water vapor. Eq 33 would still be true as long as wv was at saturation everywhere, but it seems odd to make that requirement.
34 seems to be specific for saturated adiabatic ascent
Well, 34 is used in deriving 37, which is then applied to a Hadley Cell. But my main query here was Nv, which is defined as saturated wv density, but used in places like 33, where I would have expected to see just wv density.
Yes . That s what I think too .
As they explain below 34 that dNv/dx = 0 because they have horizontal isothermal layers, I am pretty sure that 34 comes just by considering saturated ascent.
It is clearly independent from 32 and 33 but it is necessary to derive 37.
It is surely not so complicated but it needs some time that I have not rght now.
If you take dNd/dx=0 and dNv/dx=0, then my expression derived from 32 and 33 is almost exactly 34, but with Nd replacing N.
But if 34 is limited to saturated adiabatic ascent with no horizontal density gradient, then that must greatly restrict the application of 37. Can it be used for Hadley cells?
Nick, my reply to your derivation of an equation that you characterize as almost exactly Eq. 37 was given here.
The physics of Eq. 34 was discussed in greater detail here.
Anastassia,
I am referring to the expression
S = u.( d Nv/dx – Nv/Nd d Nd/dx) + w.( d Nv/dz – Nv/Nd d Nd/dz)
in #62. That is obtained simply by algebraically eliminating velocity derivatives from eq’s 32 and 33, with no extra assumptions. If you set the x derivatives of Nv and Nd to zero, you then get
S = w.( d Nv/dz – Nv/Nd d Nd/dz)
which as I said is the same as your Eq 34 with Nd replacing N.
Tomas,
Ok, I think I agree now; My thought experiment wasn’t relevant since I think it would imply extra terms with dn (change in moles of air).
Also, in trying to understand the apparent conflict between section 2 and the reviewer comment – I realize now (which should have been obvious) that in each case a different thermodynamic process is assumed, so it should not be surprising that the results don’t agree. I guess the real question is which one is a more relevant approximation to a real atmospheric process, and that is not clear to me at the moment.
Tomas, thank you very much for these comments.
If you look at our team — Victor and I are from Russia, Douglas is from Northern Ireland, works in Uganda, Antionio lives in Brazil, Larry is American with Chinese roots… — we have drastically different cultural and educational backgrounds spanning a good part of global diversity. Motivated by understanding the potential importance of these findings (which each of us has come to appreciate via his own route) we did envisage the task “to explain and communicate” as a major priority. I am very happy to see from your comments that we seem to have achieved something here. It was not easy work.
Thank you for raising the hydrostatic equilibrium issue. Indeed, our use of the hydrostatic equilibrium equation begs the question: to what degree your results depend on the accuracy of this equilibrium? I dwelt on it above and I will do it again now somewhat differently.
Most part of air is dry air. If we write the equation of hydrostatic equilibrium for dry air
it is easy to see that we will obtain a more or less correct atmospheric scale height
of around 8 km:
.
However, we also have water vapor with its highly non-hydrostatic distribution and a scale height
of the order of 2.5 km:
.
For total air pressure $p = p_d + p_v$, if there is no change in the above distribution of $p_d$, the pressure distribution will be highly non-hydrostatic with vertical accelerations producing velocities of 50 m/s over atmospheric scale height. We discussed it in our HESS paper, page 1023. This shows once again that the “small”
can indeed bring about a very sizable mischief.
As I said before, total disequilibrium pressure difference would in this case be of the order of
. Apparently this is never (routinely) observed in the atmosphere. This means that the distribution of dry air is continuously modified such that the removal of vapor aloft during moist adiabatic ascent is compensated by some excess of dry air arriving from below. While hydrostatic equilibrium is never exactly observed (otherwise the atmosphere would be static), deviations from it due to this continuous adjustment in the ascending air are much smaller than 
over atmospheric height. Such that 
. This is all we need to have our result Eq. (37).
Most part of air is dry air. If we write the equation of hydrostatic equilibrium for dry air
dp_d/dz = -\rho_d g
it is easy to see that we will obtain a more or less correct atmospheric scale height h_d of around 8 km:
dp_d/dz = -p_d/h_d.
I have discovered the specific humidity at the tropopause appears to be correlated to the level of solar activity.
http://tallbloke.files.wordpress.com/2010/08/shumidity-ssn96.png
I would be very interested to know if your theory could help in understanding why that should be the case. Thank you for your very interesting paper
Thank you Judith for introducing us the very interesting paper. As a whole, I like it. It gives a new insight for understanding the climate. According to the paper, the climate system is not only a heat engine, but also a steam engine (or almost the latter). Thus, it is apparent again that “climate phenomenology” is incomplete and under development today.
I hope the idea of Makarieva et al. can be extended to give solutions to other issues of the climate system: for instance, problem of the constant relative humidity assumption, and if possible, real values of climate sensitivity.
Kiminori Itoh, Yokohama National University, Japan
Anastassia
Further few considerations.
I have been asking myself if the problem was correctly closed.
This question was triggered by the fact that u.dN/dx = S.
This admittedly puzzling expression asks the causality question – is it u (and dN/dx) that drives S or is it S that drives u and dN/dx.
Btw I find that the use of S is not a happy one even if it is (S)ink and
(S)ource. As it is a condensation/evaporation rate another symbol Re or Rc would be better.
So.
We have 6 unknowns – u, w, Nv, Nd, N, S.
We have 3 obvious independent equations – 2 times mass conservation (32&33) and N=Nv+Nd.
3 equations missing.
Clearly S plays a special role because it is a rate and because it describes a complex process – phase change.
How fast will water condense?
If one uses the isothermal layer model and supposes that a volume is saturated all the time (so no exotical oversaturations or undersaturations) then this is easy: dS=dPv/RT . w
This adds 2 unknowns (Pv and T) but for Pv I have the state equation and for T Clapeyron. I will assume that Eq 34 can be derived from here so Eq 34 IS this relation.
2 equations missing.
Those must involve u and w.
Clearly u and w can’t be independent because one sees that S depends only on u via u.dN/dx = S and at the same time only on w (via Eq 34).
Those 2 missing equations are exactly Navier Stokes because, indeed, in a 2D model like yours, you will get 2 more equations with also Ps, Ns and Us but no additional unknowns.
OK so now you can sleep well because your problem is closed :)
Obviously because you didn’t use Navier Stokes, you treated a non closed problem.
The non closure manifests itself by the fact that you have a w/u factor which is not constrained and can be anything you want.
You can empirically estimate what this can be but only over large scales or with additional hazardous (and unnecessary !) assumptions.
I am not sure that large scale averaging is very relevant to processes that are per definition extremely local (they happen inside clouds).
And if you take only different kinds of clouds, then you see that the values of w/u have such a huge dispersion that an average is useless.
As this has been too long already, just one last comment.
Of course one can obtain intermediary results even in a non closed problem.
Your Eq 37 is such an intermediary result and would be perfectly valid if 34 is correctly derived.
But it is only by a correct closure (in this case Navier Stokes) that you will be able to derive right values of S, u and w which can then be compared to empirical evidence.
“u.dN/dx = S”
Yes, this is what you get when you put eq 36 and 34 together. And it does seem to indicate that the maths has gone astray. I couldn’t reproduce the deduction of either 35 or 36 from the stated equations. I think this needs to be written out in more detail.
But u.dN/dx = S? That could give a finite condensation rate even from dry air.
No. Obviously Nv is non zero (even saturated) by definition so these equations don’t apply to dry air.
What this means is that there must be a horizontal movement (u non zero) and a horizontal pressure gradient (dN/dx non zero) in order for condensation to take place.
I don’t find this shocking.
But I would definitely like that it be clearly derived.
I can’t help feeling uncomfortable when one deals with a non closed problem – 4 equations for 6 variables is 2 too few.
There is much potential for implicit hypothesis or a circular reasoning to sneak in.
It could be very cold air. Anyway dN_v/dx was assumed zero, so it’s actually
u dN_d/dx = S
which means, for example, S independent of temperature.
No it doesn’t mean that, you are making simple things look confused.
It’s 34 that governs condensation.
This only means that u.dp/dx is some function of z only and this is not shocking for a 2D isothermal layer model.
The only problem is that a non closed problem has an infinity of solutions so there are (almost) necessarily some hypothesis to add constraints. 2 in this case.
I had thought that u and w were assumed to be determined by some external solution process. For even if you added, say, Navier-Stokes equations, I don’t see how you could get boundary conditions for this type of problem geometry.
Oh I see.
It doesn’t matter. As this deals with infinite horizontal planes, you’d use f.ex the no slip condition. The only “logical” surface available here is a plane.
But I didn’t really mean to solve N-S explicitely or even the 6 PDE system of Anastassia.
I thought more on closing the problem and then derive equations like 36 and others without ambiguity.
There are tons of ways to simplify 2D N-S and if I was doing climate research, I’d try to tackle the closed problem.
But I am not so won’t.
This is a good point. Actually as far as we have a sink proportional to vertical velocity, and velocities are ultimately determined by the pressure gradients produced by this sink, the question arises: what determines the magnitude of this sink?
I will talk more on this in the post on scale analysis, but here just to mention that for large-scale steady-state global pattern like Hadley cell the value of vertical velocity w, as well as the magnitude of sink S (precipitation), is apparently bounded from above by the solar power.
Anastassia, a key issue is where your w comes from. Not obvious given your hydrostatic assumption.
Judy, here and here I explained that we only need a very approximate hydrostatic equilibrium for Eq. 37 to hold. This allows for a wide range of w.
Thank you for this discussion. A few thoughts. The bottom line is that to solve the system at all in the presence of a mass sink, we need an independent equation (our Eq. 34). I am not aware of any theoretical research on this topic or any other expression that could rival our Eq. (34) for moist adiabatic ascent.
Mass sink in a unit volume in the steady state invariably involves contraction of the ambient environment to keep the local pressure constant. This contraction is equivalent to work performed on the considered air volume. Therefore, apart from being an intermediary result, our estimates for Hadley cell describe the real power of this process.
Mass sink produces a pressure gradient, which carries a certain store of potential energy. Potential energy is not accounted for in the 1st law of thermodynamics and cannot be estimated from it. Therefore, our Eq. 37 represents a peculiar pressure gradient force that makes an account of this potential energy release. Without such an account, solution of the equations of hydrodynamics will not satisfy the energy conservation law.
I agree and I recognize that 34 must be independently derived.
However it would be helpful if you could write in a few lines how you derive 35.
I can’t do that because the problem is not closed.
At least it appears (unless I am very mistaken) that 35 does not follow by virtue of 32,33, 34 alone.
It is very important as without 35 there is no way to derive 37.
We put Eq. (34) into Eq. (33) to obtain (using dNv/dx = 0):
.
(I will use d instead of partial d):
Canceling wdN_v/dz in both sides and multiplying all the terms by
we have:
,
which is Eq. 35.
OK that’s clear, thanks .
Then 36 and 37 follow easily .
So indeed the key is to derive independently 34 (this part is not yet so clear) and once you have that, 35 , 36 and 37 follow.
How do you interpret that S = u.dN/dx which is also a consequence of 35?
S = u.dN/dx is the central result, it is very clear: the more intense mass sink S is, the greater density (and pressure) (dN/dx) gradient it will produce at a given u.
For a given S, the greater the value of u, the sooner the mass disturbance is relaxed, such that the steady state density gradient becomes smaller.
OK that’s what I thought and what I also said to Nick.
34 drives condensation and condensation drives dp/dx.
So that leaves for me now only the 34 and the hydrostatic assumption.
If I have time I will try to come at 34 from another angle.
Just in case, I list the relevant links in one place for convenience:
Hydrostatic approximation was discussed here; the physics of Eq. 34 was discussed here. I also discussed how it differs from the derivation in Curry and Webster (1999) here.
Tomas, let me share some further thoughts on the place of Eq. 37 in the general system of the equations of hydrodynamics.
Energy can be imparted to gas in two ways — by a force (if we push the gas) or by a heat flow. The action of forces is taken into account in the Euler (Navier-Stokes) equations, while the heat flow is accounted for in the 1st law of thermodynamics.
For a non-condensing gas with conserved mass we have a closed system of equations. We now introduce a new process, a mass sink. To introduce it correctly, we need to do three things (1) to account for the mass loss in the continuity equation; (2) to account for the heat fluxes that accompany this mass sink in the 1st law of thermodynamics and (3) to account for the volume-specific forces that are brought into action by the existence of this sink.
Action (1) involves specifying S in the continuity equation. Action (2) involves accounting for latent heat release in the 1st law of thermodynamics. Action (3) involves estimating the volume-specific force associated with this sink, to be put in the list of volume-specific forces in the right-hand side of the NS equations.
Our main point is that in atmospheric theory people first performed Action (2) (latent heat was accounted for long ago), then, quite recently, they put an explicit non-zero term into the continuity equation (performed Action 1), but they DID NOT estimate the pressure-gradient force associated with the mass sink, i.e. they did not perform Action 3. Therefore, the existing accounts of mass sink violate the energy conservation law by ignoring the work of the pressure gradient force, Eq. 37, associated with such a sink. As Judy pointed out, for this reason there is no way of “testing” the effect by the existing models — they are missing the very process that they would be supposed to test.
Our work is devoted to describing this pressure gradient force. It is not a solution of the problem as a whole. It identifies a force that has been overlooked in the NS equations. Eq. 37 should be considered as such. And the basic estimates that we perform show that this force is the major force in atmospheric dynamics.
Anastassia
Our work is devoted to describing this pressure gradient force. It is not a solution of the problem as a whole. It identifies a force that has been overlooked in the NS equations. Eq. 37 should be considered as such.
This view is not correct .
N-S is not about forces .
It is about local momentum conservation – you plug in any forces you want or need, N-S can’t “overlook” anything because it makes sure that momentum is conserved.
As I explained I think quite clearly above , your system of equations is not closed . You miss 2 equations for 2 unknowns .
And be sure that it is not a coincidence that N-S provides exactly the 2 more independent equations that are needed for closure .
It seems also obvious to me that it must be so – one cannot solve a dynamical problem just with mass continuity and energy conservation .
Even if one accepts 34 from where everything follows , in 37 you have a free LOCAL parameter w/u which can be anything you want . The range over which it may vary is simply huge .
You are not allowed to expand 37 which is a local equation over large distances and you are also not allowed to take “averages” of w/u .
That’s why I repeat one more time – if the system of equations you propose in your paper should be closed in order to conclude on a dynamical result which could be confronted to evidence, you simply need 2 more equations.
Without that necessary closure, 37 makes only predictions about a local result in a differential volume dV .
It can’t be integrated because you don’t know what w/u does .
And doing local experiments inside clouds to check the theory wil be notoriously difficult .
Tomas, actually I agree with all that you say. I’ve been thinking my previous comment was not particularly lucky and did not carry my message properly or accurately.
The point is that as we have obtained an expression for the pressure gradient force, we can check its validity using the observed values u and w. We do know them. We can check the relevance of this force for atmospheric motions. True, w/u can vary over a huge range, but we can ask what the force will be for a given value of this ratio. And we can see that what Eq. (37) predicts conforms well to observations (both for Hadley cell and for hurricanes).
The question of spatial scale where Eq. 37 is valid is a separate issue. As I said before, on a scale of a single cloud it will not be valid.
There is another interesting discussion going on the Air Vent, Weight of Water and Wind, Hurricane Pro’s Weigh In.
Tomas,
It can be integrated over those regions where w/u is observed to change sufficiently slowly and can be approximated by an average value, which is close to the case of Hadley cell. Yes, right now we do not predict that the circulation will be such as it is, but given that it already exists, we point out that this force is apparently a major one there.
Tomas,
Just to point out that here a nice and simple formula for hurricane power is provided based on the condensation-induced mass removal. This is a different approach for illustrating the relevance of mass removal that is based on consideration of potential energy release. It yields essentially the same results.
Dear Anastassia,
It’s refreshing to see your paper discussed and the immensely positive interaction it generates. Have you had a chance to read my comments addressed to you via email?
Thank you,
Marc
Dear Marc
Thank you for your comments and email. You suggest that we should have cited the work of Dr. M. Leroix. I consider this work very interesting and informative. However, Dr. Leroix ascribes the driving force behing atmospheric movement to the thermal deficit of the poles, which brings about mobile polar highs. In our work, we are advancing the idea that the major driving force behind atmospheric motion is not connected to either thermal deficit or excess and we do not criticize or overview in any other way previous work on this subject.
My apologies for spelling mistakes, which I am struggling with all the way. I meant the work of Professor Marcel Leroux.
Dear Anastassia,
For clarity and the benefit of readers, I shall resume the points I raised regarding Professor Leroux’s work and why I believe a reference would not have been out of place. I understand your answer was articulated from a physicist viewpoint. If on a theoretical causation the two works differ notably, my naturalist remarks were based on the observation wealth Prof. Leroux founded his description of facts.
In your M10 abstract, you write:
“Building from fundamental physical principles we show that condensation is associated with a decline in air pressure in the lower atmosphere. This decline occurs up to a certain height, which ranges from 3 to 4 km for surface temperatures from 10 to 30 C. We then estimate the horizontal pressure differences associated with water vapor condensation and find that these are comparable in magnitude with the pressure differences driving observed circulation patterns.”
I noted that, not dissimilarly to Prof. Leroux and thus in contrast with the preferred meteorological theories giving the altitude jets preeminence, you attribute a very significant importance to the lower layers. Thus a reference to his work would seem appropriate in this context. If anything, Leroux’s description of these exchanges geometry is based on thorough observation and one would therefore expect a new physics theory driving circulation patterns to reconcile Leroux’s observations even if the driving mechanism differs. This geometric reality doesn’t change.
As an aside to this, two supplemental points:
You write in Dr. Curry’s blog:
“For example, Hadley cells (the major feature of Earth’s atmospheric circulation) are driven…”
And in the Air Vent:
“By the way, regarding heating. This summer in Russia there was anomalous heat over a thousand kilometers. This heat-struck territory was some 10 degrees (!) warmer than the neighboring regions. One could expect the warm air to ascend, after all. When, if not now, people thought. But what do you think? The air descended instead, and did so for two months, while the rising air flow somehow appeared in the neighboring colder regions in Europe that got flooded.”
Leroux’s synoptic observations consigned in “Meteorology and Climate of Tropical Africa” show the Hadley cells only exist partially due to the significant Trade Inversion discontinuity. Therefore of the three usual “cells” traditionally invoked in the general circulation, Leroux demonstrates that only one and only a portion of it really exists.
As for surface super-heated air not ascending under Anticyclone Agglutination conditions, Leroux in “Dynamic Analysis of Weather and Climate” page 70, paragraph 3.5 explains that as the moisture level plummets under the same strong insolation, the surface ultra heated air does not rise because of the high pressure anticyclone, constantly renewed by Mobile Polar Highs. Satellite animations show that despite Anticyclone Agglutination perceived stability on the statistical level, in fact on the synoptic scale AA are constantly replenished by an influx of MPHs coming from the north. The pulsations lines were very clear and allow anyone to follow the evolution of this lower layer features. It was not any subsidence of cold air from above.
I trust these remarks helped framing the reason why a Leroux reference and the associated knowledge in meteorology would have been pertinent to the naturalist side of the discussion.
Cordially,
Having spent some time at RealClimate, I would say that the characterization of “just fine” to describe Gavin’s feelings about the models is accurate. Considering that many of these models haven’t been falsified only due to their youth and their generous error bands, I would say that they are not “just fine”. Certainly a question about what would happen if they missed some major factors effecting climate does not deserve to be treated with the contempt of comparing it to a question about the moon being made of green cheese. I think that papers like Susan Solomon’s about water vapor changes in the stratospher show us that there is still much to learn about the climate and that these things have significant effects that are not modeled. Svensmark’s cosmic ray research may end up being another factor that will force significant changes on the models. One would think that Gavin’s arrogance about the models would be curbed by the current 12 year climate flat trend. He hand waves away the flat trend by attributing it to natural variation. But if you ask him to identify the elements of natural variation that are responsible, he runs away from the question. You can be sure that if Gavin cannot identify those elements, then he is not covering them in his model.
On the point of pressure decrease, I think that condensation would lead to a pressure drop in the column below the condensation that could be communicated to the ground.
But that is not the point, in general what is the “speed limit” for vertical pressure waves in a GCM? There would appear t be a limit set by a layers depth divided by timestep. So a ~300m layer would need updating at least once every second for a compression wave.
Mind you I have no idea if this would matter, I’m just interested to see if I understand the range of reproducible velocities correctly.
Or do the layer boundaries move instead of air through the boundaries?
Alex
I received a very helpful and relevant email message from a climate modeler at GFDL, here is the essence of the message:
“[It] was brought to my attention your blog article and that there are
related on-going debates about the worthiness of today’s climate models
due to the perceived lack of *precipitation as mass sink* effect. I just
want to point out that this effect was correct implemented in GFDL’s
CM2.1 (which was used in IPCC AR4) and all models developed at GFDL
since. In fact, the NASA GEOS-4 and GEOS-5 GCMs also shared the same
attribute (because we used the same FV dynamical core; this core is also
being used at NCAR for AR5 experiments).
We can get into more technical details if you wish. But I would like to
point out that the “vertically Lagrangian control-volume discretization”
used the the finite-volume dynamical core (see4 Lin 2004, Mon. Wea.
Rev.) allows us to have a local mass sink/source due to moisture
changes. All other climate models that I know can’t accomplish this
because the vertical coordinate is tied directly to surface pressure.
I did some sensitivity experiments with and without
this “moist” mass-sink effect, and we did not find a strong positive
feedback mechanism we suspected. This is perhaps due to our experiments
design in which we prescribed the SST. Nevertheless, it is important to
have this effect for correct mass conservation and also for accuracy reason.”
I would like to point out that a very similar claim has been recently put forward by Dr. Emanuel (see tAV):
This brought up an intense discussion with a rather detailed analysis of the propositions made during reaching these conclusions. The entire thread is rather heavy (over 430 comments). So I point out here, here and here why I think the performed analysis was incorrect. See also here for my explanation why the model results nevertheless conformed to observations.
In the paper of Lin (2004) that is referred to in the comment there is no account of a mass sink (see Eq. A4 on p. 2306). If there are published sensitivity studies of a mass sink, it would be very valuable if a reference is provided, such that the physical assumptions underlying the analysis could be scrutinized.
A statement made earlier by AM suggested that industrial carbon emissions are ‘irrelevant’. Perhaps, though unlikely. How does this claim by AM measure up with the evidence for ocean acidification? Please sidetrack for as little time as required. I would like to here from AM in particular.
Dear All
You may be interested that my M10 co-authors and myself just posted two comments at the M10 review website. These address some common concerns.
In these comments we:
(a) show why previous model based studies of the precipitation mass sink have incorrectly gauged the magnitude and importance of condensation driven pressure gradients;
(b) clarify what eq. 34 (in M10) represents by examining the vertical distribution vapor and dry air in a column of moist air at equilibrium. (There will be more on this soon as we are also drafting our response to Nick Stokes.)
You can find the comments here
Best wishes
Some of the readers may still be interested the fate of our draft paper “Where do winds come from?” Several substantive comments and replies, including our reply to the open review by Dr. Curry, have been posted in the discussion. We believe that responding to these comments has clarified our presentation and helped identify and address some misunderstandings. We are grateful to all those who have contributed.
A second referee was appointed –- many thanks to all of those who helped with that. Now we wait for their judgements. The open discussion was extended until at least April 7th, 2011 (from the original deadline of December 10th, 2010).
We are glad to be part of this new way of assessing and refining science and invite you to join! We welcome comments, criticisms and suggestions. How can we be more convincing? (Additional developments on this topic can also be found here.)
Hi Anastassia, thanks for the update