Guest essay by Eric Worrall
Admitting some cloud error is as close as most climate modellers come to admitting their projections are not fit for purpose. Note the image above is from Pat Frank’s paper about cloud error, not Paulo Ceppi and Ric Williams’ paper. More below.
Why clouds are the missing piece in the climate change puzzle
September 11, 2020 8.01pm AEST
Paulo Ceppi Lecturer in Climate Science, Imperial College London
Ric Williams Professor of Ocean Sciences, University of LiverpoolHow much our world will warm this century depends on the actions we take in coming decades. In order to keep global temperature rise below 1.5°C and avoid dangerous levels of warming, governments need to know how much carbon they can emit, and over what timeframe.
But current climate models don’t agree on where that threshold lies. In new research, we discovered one of the reasons why there is such a large range of estimates for how much carbon can be safely emitted: the uncertain behaviour of clouds. In some climate models, clouds strongly amplify warming. In others, they have a neutral effect or even dampen warming slightly. So why are clouds likely to play such a pivotal role in deciding our fate?
…
Clouds can act like a parasol, cooling the Earth by reflecting sunlight away from the planet’s surface and back into space. But they can also act like an insulating blanket, warming the Earth by preventing some of the heat in our atmosphere from escaping into space as infrared radiation. This “blanket” effect is particularly noticeable during the winter, when cloudy nights are typically much warmer than cloud-free ones.
…
While we do know that clouds will likely amplify global warming, there is still a great deal of uncertainty about how strong this effect will be. Here climate models are of little help, as they can only simulate the bulk properties of the atmosphere over scales of tens of kilometres and several hours. Tiny cloud droplets form and evaporate in minutes. Models miss these small-scale details, but they’re needed for accurate predictions.
Climate models have to resort to simplifications in order to represent clouds, which introduces error. As different models make different simplifications in their portrayal of cloud processes, they also make different predictions of the cloud feedback, which results in a range of global warming projections and differences in our remaining carbon budget. For a given future carbon emissions scenario, clouds are the single most important factor behind the differences in future warming predicted between models.
…
Read more: https://theconversation.com/why-clouds-are-the-missing-piece-in-the-climate-change-puzzle-140812
The abstract of the study;
Controls of the transient climate response to emissions by physical feedbacks, heat uptake and carbon cycling
Richard G Williams1,4, Paulo Ceppi2 and Anna Katavouta1,3
Published 11 September 2020 • © 2020 The Author(s).
Published by IOP Publishing LtdThe surface warming response to carbon emissions is diagnosed using a suite of Earth system models, 9 CMIP6 and 7 CMIP5, following an annual 1% rise in atmospheric CO2 over 140 years. This surface warming response defines a climate metric, the Transient Climate Response to cumulative carbon Emissions (TCRE), which is important in estimating how much carbon may be emitted to avoid dangerous climate. The processes controlling these intermodel differences in the TCRE are revealed by defining the TCRE in terms of a product of three dependences: the surface warming dependence on radiative forcing (including the effects of physical climate feedbacks and planetary heat uptake), the radiative forcing dependence on changes in atmospheric carbon and the airborne fraction. Intermodel differences in the TCRE are mainly controlled by the thermal response involving the surface warming dependence on radiative forcing, which arise through large differences in physical climate feedbacks that are only partly compensated by smaller differences in ocean heat uptake. The other contributions to the TCRE from the radiative forcing and carbon responses are of comparable importance to the contribution from the thermal response on timescales of 50 years and longer for our subset of CMIP5 models and 100 years and longer for our subset of CMIP6 models. Hence, providing tighter constraints on how much carbon may be emitted based on the TCRE requires providing tighter bounds for estimates of the physical climate feedbacks, particularly from clouds, as well as to a lesser extent for the other contributions from the rate of ocean heat uptake, and the terrestrial and ocean cycling of carbon.
Read more: https://iopscience.iop.org/article/10.1088/1748-9326/ab97c9
The authors assert that if we had a better understanding clouds, the spread of model predictions could be reduced. But there is some controversy about how badly cloud errors affect model predictions, and that controversy is not just limited to climate alarmists.
Pat Frank, who produced the diagram at the top of the page in his paper “Propagation of Error and the Reliability of Global Air Temperature Projections“, argues that climate models are unphysical and utterly unreliable, because they contain known model cloud physics errors so large the impact of the errors dwarfs the effect of rising CO2. My understanding is Pat believes large climate model physics errors have been hidden away via a dubious tuning process, which adds even more errors to coerce climate models into matching past temperature observations, without fixing the original errors.
Climate skeptic Dr. Roy Spencer disagrees with Pat Frank; Dr. Spencer suggests the cloud error biases hilighted by Pat Frank are cancelled out by other biases, resulting in a stable top of atmosphere radiative balance. Dr. Spencer makes it clear that he also does not trust climate model projections, though for different reasons to Pat Frank.
Other climate scientists like the authors of the study above, Paulo Ceppi and Ric Williams, pop up from time to time and suggest that clouds are a significant problem, though Paulo and Ric’s estimate of the scale of the problem appears to be well short of Pat Frank’s estimate.
Whoever is right, I think what is abundantly clear is the science is far from settled.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Wow, the atmospheric CO2 must be over 1000 ppm. If you add 1% to the CO2 each year, you would use the same calculation you use to calculate compound interest. The fact that blooper escaped all the authors and the editor and the peer reviewers speaks very poorly of the mathematical knowledge of the climate community as a whole.
Yes, confirming a 1%/year rise starting from the level of CO2 that existed 140 years ago would be:
280 ppm*(1.01)^140 = 1,128 ppm . . . and that would be ± uncertainty that GLOBAL atmospheric CO2 concentration was actually 280 ppmv back then.
They are so used to qualitative arguments that when the attempt to make a quantitative statement, they’re unable to do so.
I guess they meant “annual rises in atmospheric CO2 over 140 years. The current rate of increase is about 1% of the original concentration.”
Then the English majors got involved.
Common sense suggests to me that increasing clouds would result in cooler days and warmer nights. Overall, these are good things.
When heat kills it is usually the midday heat under the sun. When cold kills, usually it is the nighttime lows that are worst.
So what if overall global temps go up a degree or two, as long as peak temps are moderated by several degrees. And more clouds means more rain. Ask California how nice that would be.
Would a post or even an article on turbulence by Christopher Essex be helpful here?(Enter the Navier Stokes equations!)
In order for clouds to form, the water needs something physical to condense on, e.g. dust. However, dust is not the only thing on which water will condense.
Has anyone ever heard of that classic piece of demonstration apparatus known as the “Wilson Cloud Chamber,” or seen one in action? I have. Saw one in a physics demonstration lab way back about 1955, and have remembered it ever since.
As SpaceWeather.com has noted, cosmic rays are increasing. It seems to me that outside of the West Coast I have been seeing a bit of cold weather lately, cold weather which seems to escape the notice of the news media for some reason or other.
The next decade could be very interesting. If some unconsidered factor such as cosmic ray influence on clouds is responsible for the model cloud error identified by Pat Frank, a substantial change in that unconsidered factor could throw models even further off.
Peter W
Yes, almost any nuclei will do. So, that includes salt from the oceans, various minerals that are in common wind-blown dust, volcanic ash, micrometeorites, and even bacteria. The last two are rarely ever considered. However, they might vary naturally over time, giving rise to variations in cloud cover. The science is NOT settled!
You spend $2T to lower the temperature by 1°C via CO2 and the average cloud cover changes by a few % and wipes that all out. You don’t have to be an Einstein to see how stupid that is. H2O is significant, CO2 is not.
$2T for 1C?
I think that the Green New Deal is supposed to run us about $100T and that’s supposed to cut us back from 2C to 1.5C.
More like $200T for 1C that could be thwarted by a small decrease in cloud cover.
Why do we still have so many models?
If we have a consensus that the science is settled, we should only have one model.
Yep.
We must have diversity. And also smoke, and mirrors.
The keys to the hydrological cycle are the fixed points of the change of state. If these were at different points as they are for different elements then there would be a totally different environment.
I don’t agree with Dr. Spencer. Random errors may cancel. Uncertainty does *not*. Uncertainties add, they don’t cancel.
The issue with clouds is not random error. It is an uncertainty about how clouds act. There is no probability density that is associated with random error when it comes to clouds. There is jut plain uncertainty and uncertainty has no probability density, it is not a random variable and, therefore, cannot cancel.
Every time a climate model is iterated with an uncertain cloud impact the uncertainty of the combined impacts grows. After some number of iterations the uncertainty becomes larger than what the model predicts.
You can’t make the models better by just ignoring the uncertainties built into the model. The uncertainties don’t just go away.
My understanding of what Dr. Spencer was saying is that the Stefan-Boltzmann effect dominates – heat makes it out one way or another, so errors in ascribing heat transport to any given mechanism don’t have a major impact on the overall outcome. Dr. Spencer also suggested a while ago that our cloud observations are not good enough to ascribe a cloud error the way Pat Frank has done.
Having said that, I’ve got to say I find Pat Frank’s argument about propagated model error compelling, but I’m strictly an observer in this debate – I’m not about to try to go toe to toe with someone with Roy Spencer’s expertise!
Roy Spencer’s argument is misguided because all the models are in global energy balance, at the same time that they are wrong about clouds and thus wrong about longwave cloud forcing.
The fact that errors are adjusted so as to cancel out does not mean the model can make correct predictions.
Uncertainty in energy is not an energy. It’s a statistic that tells you how accurately you know that energy.
Many here understand that distinction. Every studious second year engineering, physics, and chemistry undergraduate understands it
Somehow, that distinction is opaque to climate modelers.
I get what you are saying Pat, propagating known errors through the calculation to determine their possible impact on results is something I learned in high school (I had good teachers). It seems a really obvious way to test the value of a model prediction.
As far as I can tell Roy Spencer’s objection hinges on the idea that clouds are an effect rather than a cause, so changes in cloudiness don’t really affect the overall energy flow or surface temperature, but I could be wrong – I’m not 100% confident I understand Roy Spencer’s response.
Eric,
Uncertainty is not a known “error”. Uncertainty doesn’t have an impact on results, it just makes the result uncertain. If the inputs are uncertain then the output is uncertain. That is not the same thing as an error.
It doesn’t matter if clouds are an effect, they are still a cause as well. Clouds have an impact on the albedo of the earth. They do reflect energy away from the Earth. If they reflect energy away from the earth then they must be accounted for in the energy in/energy out calculations. If the amount of that energy reflection is uncertain then the final value of the calculations are uncertain as well.
Thanks Tim, I should have said “uncertainty”.
I had a conversation with Roy Spencer a while ago. I’m having trouble accepting that Roy Spencer doesn’t understand propagating uncertainty through the calculation, which makes me doubt my understanding of Roy Spencer’s response. What I would really like to see is another detailed explanation from Roy about why he thinks Pat is wrong.
Dr. Spencer isn’t coming out of left field. It appears to me that he is simply applying what I was taught in validating a model. That is, you do a hindcast applying the model to the data set over the period for which you have data. If the model agrees reasonably well with the historical data then that agreement constitutes validation. Dr. Spencer’s argument seems to me is that the hindcast doesn’t show the kind of geometric growth in uncertainty (measured in the hindcast by the difference between hindcast model values and data values!!!) that Dr. Frank’s paper explains.
This leads to a cognitive dissonance. It seems to me that in Dr. Spencer’s view, Dr. Frank can’t be right because the hindcast doesn’t show that large an uncertainty nor geometric growth. Uncertainty is being evaluated with different metrics, however. Dr. Frank is estimating uncertainty using propagation of uncertainty based on known uncertainties in cloud modeling. Dr. Spencer is evaluating uncertainty based on the fit between model hindcasts of historical values and historical data.
Which is correct? Dr. Frank’s evaluation of uncertainty is well grounded. Dr. Spencer’s evaluation of uncertainty is clouded by the “tuning” of the models which dampens uncertainty within the calibration range. Please refer to Dr. Frank’s rebuttal on October 15, 2019 at https://wattsupwiththat.com/2019/10/15/why-roy-spencers-criticism-is-wrong/. Please refer to his Figure 4: RCP8.5 projections from four CMIP5 models at
?resize=384%2C330&ssl=1. Up until about the year 2050, the models mostly agree with each other (which doesn’t mean they are right!). After that, they start to diverge dramatically. I believe the agreement between models in the early period is an artifice of the “tuning.” After the “tuning” wears off, model divergence reflect the propagation of uncertainty that Dr. Frank has shown in his paper.
The same thing is true of forecasts of hurricane tracks. These are sometimes referred to as “spaghetti” graphs. There is agreement among models followed by large divergences. I reiterate my point that the usefullness of these models is limited to short term forecasts.
Phil,
Matching hindcast data is really meaningless. A plethora of different combinations of math equations can be created to do data matching. None of the combinations need to be based on the real world. All the models eventually devolve to a linear equation with different slopes. That alone is concerning. Is the future temp gong to be that predictable?
Eric,
Pat is not wrong. Nothing from Mr. Spencer will change that. It simply doesn’t matter the approach you use in the climate models, be it a radiative balance, a thermal balance, or something else. Clouds *should* be a major contributor to any model and if the cloud factor is uncertain then the output of any model will be uncertain. If the cloud factor is uncertain at the start then the output of the model grows in uncertainty with every iteration. Sooner or later the uncertainty overtakes the ability to define differences in the output of the model.
I am not surprised at anyone today not understanding how to propagate uncertainty. It seems to be a subject that is not taught or observed very often any more. Far too many people don’t even understand the difference between dependent measurements (measuring the same thing multiple times with the same measurement device) and independent measurements (measuring different things with different measurement devices). One is error and is subject to the law of large numbers and the other is uncertainty and is not subject to the law of large numbers.
Suppose you have a disk you want to calculate the inertia of. The formula is 1/2 x m x r**2. You have two quantities to measure, mass and radius. Like with temperature you only get one measurement of each. Your measurement of the mass and the radius will have an uncertainty factor. How do you combine the uncertainty of each to come up with the uncertainty of of the inertia? You do it with root-sum-square. Just like Pat has used in all of his answers and writings.
I don’t care what Mr. Spencer says, every factor he uses has an uncertainty. And he only gets one shot at each measurement so each measurement is independent not dependent. So the uncertainties add root-sum-square. The only defense Mr. Spencer can have is that everything he measures is 100% accurate. I simply don’t believe that and never will.
The water vapor vs temperature relationship is not linear.
Increasing temperature greatly increases water vapor which condenses into aerosols (clouds).
More clouds reflect more sunlight.
There is nothing uncetain about clouds increasing global albedo.
Increasing global albedo is cooling.
Any 8 year old can see clouds build up during the day. Then dis-appear at night.
“Why clouds are the missing piece in the climate change puzzle”
This statement suggests that everything else is known. I would suggest that less than 20 percent is actually known to fully understand the mechanisms of medium to long term climate change.
This is an interesting catch from NoTricksZone:
https://notrickszone.com/2020/09/11/austrian-analyst-things-with-greenhouse-effect-ghe-arent-adding-up-something-totally-wrong/
An Austrian scientist finds observation data showing a positive correlation if cloud cover with warmer temperature, in a part of the world (North Pacific, Aleutian Islands) where climate models predict the opposite – a strongly negative cloud radiation effect.
They can’t both be right.
Williams et al abstract starts “The surface warming response …. ” It does not say “The total atmospheric warming response …”
If clouds intercept radiation travelling from surface to sky and make it warmer below them, that process must make it cooler above the clouds, so the total column has no change. Clouds cannot increase total energy.
Some researchers seem to think that this tiny effect on near-ground temperatures means they can say silly things like “global warming” and “the Earth has a fever.”
Why does it get so hard to make people see the obvious?
Geoff S
Can’t get clouds right. Can’t get precipitation right. Can’t get temps right on a continental or regional scale.
But if you add the garbage results on top of garbage results, you get something reasonable for a global temperature anomaly. Therefore the model’s are “close.” Only in climate science would this fly.
As I like to say, climate science is the only field of human endeavor, where you can average a bunch of wrong answers in order to find the right answer.
pat’s wrong.
next.
Steven Mosher is wrong.
Next.
Mosher
Your arrogance knows no bounds, unlike the uncertainty of sequential calculations.
Mosh is off his meds again .. next
Mosh deeps dives into his pit of IRRELEVANCE. !
Home sweet home for you, mosh !
Why is it that people who’ve studied English at university think that they are science experts? I say studied because Roger Harrabin, another English student seems rather shy about what qualification he got at the end of it. I think that most journalists call into this category. These people have no understanding of how science works and should keep their mouths firmly shut.
PS. Pay Frank is spot on.
Well look at that. No spelling errors again! ⭐️
And this time, punctuation! Super job! ⭐️⭐️
Please review the hand-out on capitalization (big letters) 🦄
Punctuation, but is it correct punctuation?
Is “next” enough of a sentence to get a period after it?
Oh Mark. Please don’t discourage the learner. He found the period key twice and even the apostrophe. The apostrophe is very advanced. There could be a comma in the future.
And you don’t even credit him for getting the spelling right on three out of three monosyllabic words.
Have you read and understood the BIPM Guide to the expression of uncertainty in measurement?
Steve, thank you for once again admitting that you can’t refute the argument.
Steve, you’ve never displayed an iota of scientific understanding. Not in the knowledge, and not in the method of thinking. Your opinion is meritless.
If any readers here have a Twitter account, please tweet my paper, here.
That seems to be the only way, apart from WUWT, that word of it gets out. Thanks, all. 🙂
It does not seem logical that the uncertainty of a model forecast could exceed the bounds of a system, as some have interpreted Dr. Frank’s paper. I was taught to calculate uncertainty separately from model forecasts. It is this theoretical uncertainty that can exceed system bounds.
When the theoretical uncertainty exceeds system bounds after a certain number of iterative forecasting steps, then the model becomes useless for forecasts beyond that time horizon. That the uncertainty range exceeds the bounds of the system does not mean that future values are expected to exceed the bounds of the system.
Another way to look at it is that the point in the forecasts at which the theoretical uncertainty exceeds the bounds of the system marks the outermost forecast horizon of the model. IIRC, most weather models have a forecast horizon of days, not weeks. The accuracy of longer weather forecasts seems to drop pretty fast the further out the forecast is. For example, you don’t see anyone forecasting where a hurricane is going to be two weeks ahead very often (think “Sharpiegate”). Accurately forecasting where hurricanes are headed just days ahead is difficult enough.
There is an unfortunate circular reasoning in using forecasts of climate models to show that the forecast uncertainty is constrained over the period for which there is data. As a first test, that just means that the models pass a reasonableness test. If model forecasts exceed system bounds during the period for which there is data, then such models are immediately determined to be invalid.
A further confusion arises from the practice of “tuning” models. I was taught to build a model and then to tune it by using smoothing constants and minimizing the sum of the squared errors between the model values and the data values over the period for which there is data by varying each smoothing constant over a certain range. I assume something similar in concept is done with climate models.
The problem with this practice is that the optimization of the tuning parameters is only valid over the calibration range. Forecasts are necessarily made outside of the calibration range. Tuning a model artificially constrains uncertainty within the calibration range. Tuning is a heuristic. It has no physical meaning. Thus the model becomes heuristic and any claim that the actual physics is being represented mathematically becomes more difficult to make.
Then a further assumption is made that the calibration done over the period for which there is data is valid outside of the calibration range. That assumption becomes less and less valid the longer the forecast is.
When using a heuristic, the modeling uncertainty is merely dampened. One cannot pretend that the modeling uncertainty has been eliminated. Thus, Dr. Spencer’s argument that the uncertainty in modeling clouds has been constrained is true only for the calibration range. Forecasts are necessarily done outside of the calibration range. Since the model is a heuristic and unphysical, the benefit of the tuning is not unlimited. Dr. Spencer is assuming that the dampening of the uncertainty during the calibration range will continue into the future as it does during the calibration range. That assumption is unsupported.
Should public policy be formulated based on heuristics? The answer is no for long term forecasts. Short term forecasts (i.e. weather) have been shown to be useful.
This seems to be the essence of Pat Frank’s argument, since the models have been coerced into agreeing with past temperatures without regard for fixing substantial underlying physics errors they don’t really tell us anything about what will happen outside their calibration period.
But Roy Spencer’s argument as far as I can tell is the cloud errors are not as significant as Pat thinks, that Stefan-Boltzmann rules the climate TOA energy balance, so if clouds are a little heavier one year, instead of the Earth retaining more heat or whatever the energy balances out via a different path.
Yet clouds must be important on some level if differences in predicted cloud behaviour are responsible for the spread of model predictions.
Glad I don’t have to figure it all out.
Thank-you, Phil. You’ve captured the whole message in a comprehensive nutshell.
All I did was put numbers in it.
Phil,
Thank you. You are among the few people who not only understand this, but know enough to write about it.
If there was a moment of awakening for me, it was in my young analytical chemistry days ca 1965-70 when the Moon missions were returning lunar rock and soil for analysis. Several labs, mostly government and university, with top reputations and credentials were engaged. After the analysis came the inevitable round robin comparisons. Whereas, for example, some labs claimed their results were accurate to +/- 10%, many times the labs differed by N times 10% where N was a number usually between 0 and 9.
Simply, the self-described performance beforehand was not achieved by the tested performance after the event. You would need a book to explain the reasons why.
Note that the analyses were done (mostly, AFAIK) independently. These days we see model runs for climate models compared with many others in a process that seems to allow rejection of those that do not look “correct”. Unfortunately, there is no way I know to determine “correct” in this modelling work.
Some of the pioneering work on this lunar soil analysis was by George H Morrison, who had the skill to dissect and report it. These days, people like Morrison was are having to run a gauntlet of opposition to their “reveals”. Geoff S
Clouds, one unknown variable. How many others?
“Climate skeptic Dr. Roy Spencer disagrees with Pat Frank”
Roy confused a calibration error statistic with an energetic perturbation. He also confused an uncertainty statistic with a physical temperature. His disagreement was wrong in every respect, and did not touch the analysis in my paper.
I’ve since found and assessed CMIP6 global average annual longwave cloud forcing calibration error. It’s a bit smaller than the CMIP5 error.
CMIP6 annual average global LWCF error is still about ±76 times larger than the annual average increase in forcing from human CO2 emissions.
There’s no way CMIP6 models can resolve the effect of such a tiny perturbation.
The last line of my paper remains true: The unavoidable conclusion is that a temperature signal from anthropogenic CO2 emissions (if any) cannot have been, nor presently can be, evidenced in climate observables.
That was the big boast of CMIP 6, that they did a better job of replicating cloud behaviour, and shock horror the hottest, most sensitive models did the best job.
Thanks Dr. Frank. I’ve been hoping for an update of your cloud error analysis ever since the alarmist buzz began around the more dire predictions of CMIP 6. I’m pleased to see that the modelers have “reduced” their error from 114x to 76x, which I think is due, at least in part, to your earlier work. Unfortunately, like squeezing a balloon, this improvement has come at the expense of even sillier predictions. Please keep up the good work.
Thanks, Frank. I was glad to find the opportunity to get resolution info about the CMIP6 models.
Also, please call me Pat. 🙂
From the article:
Yes, but that has NOTHING to do with the global warming equation. Those clouds were due to sun shining on water and evaporting it. Those clouds carry their latent heat up (in general) and the IR that holds energy under the cloud layer more effectively adds NO, ZERO new energy into the climate system. But the clouds themselve are going to reflect visible light back into space when the sun comes up, SUBTRACTING energy from the climate system. Goddammit why are climate scientist of the alarmist variety so stupid on this? It is the variability of the total cloud albedo that changes climate (and the orbital variations that cause relatively more or less albedo and angles of incidence of the light itself to alter the amount of energy entering the system on Northern Hemisphere land (as opposed to the more watery Southern Hemisphere) that makes the most profound changes.
And it says this:
Yep,, they don’t seem to have even the most basic idea how clouds get there in the first place.
Quite bizarre , really !
It’s not so simple.
They cool in the daytime. Obviously every cloud is the result of evaporation which cools the surface. But they also slow cooling in the nighttime. And you’re not mentioning that they move. They could cool a place in the daytime when forming and then blow away to a different place, where they slow down cooling at night.
People say that clouds warm us at night, implying a heat source. But what they do is keep us warm longer by slowing the rate of cooling. The cold cloud stands between us and the frigid 2-3 kelvin outer space. The clouds are much hotter than near absolute zero, so the net radiative heat loss is lower than if the cloud were not there.
Intuitively it seems that the cooling effects outweigh the heat retention effects, but can we prove that? In addition to that question, consider that even if the overall effect is overwhelmingly toward cooling, if various factors change the dispersion and persistence of cloud cover, that will modulate the amount of cooling. Less cooling has the same effect on temperature as more warming.
Rich, the most stifling uncomfortable nights are always where after a hot cloudless day heating the surface, the clouds roll in and as you say, slow the rate of cooling.
I still vividly recall how frequently these occurred as a youngster in the 1950’s because those were the restless nights where in the absence of any insulation or air-conditioning, or even a breath of wind through the wide open windows and doors, the whole family would move outside to try and get some sleep.
These days there would be few people aware of such natural occurrences given the widespread dependence on their air-conditioned insulated homes they can return to in their air-conditioned cars from their air-conditioned offices.
This is where the use of “global average temperature hides needed data. If the clouds slow warming then exactly how much do they do such? The effect of the slowed down cooling could lead to higher morning temperature (or maybe not). Do these higher morning temperatures lead to higher daytime temperatures? Maybe, maybe not. You can’tell because the use of average temps hides the data. Certainly higher minimum temps will raise the average. But if maximum temperatures are not directly dependent on the minimum temp then how do we discern that?
How do we know the less cooling has the same effect on temperature as more warming when all we have is averages?
Phil,
Thank you. You are among the few people who not only understand this, but know enough to write about it.
If there was a moment of awakening for me, it was in my young analytical chemistry days ca 1965-70 when the Moon missions were returning lunar rock and soil for analysis. Several labs, mostly government and university, with top reputations and credentials were engaged. After the analysis came the inevitable round robin comparisons. Whereas, for example, some labs claimed their results were accurate to +/- 10%, many times the labs differed by N times 10% where N was a number usually between 0 and 9.
Simply, the self-described performance beforehand was not achieved by the tested performance after the event. You would need a book to explain the reasons why.
Note that the analyses were done (mostly, AFAIK) independently. These days we see model runs for climate models compared with many others in a process that seems to allow rejection of those that do not look “correct”. Unfortunately, there is no way I know to determine “correct” in this modelling work.
Some of the pioneering work on this lunar soil analysis was by George H Morrison, who had the skill to dissect and report it. These days, people like Morrison was are having to run a gauntlet of opposition to their “reveals”. Geoff S
If you were making a climate model from scratch, the very first question that you address is, “is the climate a dissipative or a conservative system?”
https://en.m.wikipedia.org/wiki/Dissipative_system
https://en.m.wikipedia.org/wiki/Conservative_system
It’s very clear that climate is a dissipative system. This branch of theory – nonlinear thermodynamics – was developed by Ilya Prigogine. It is central to understanding climate but is ignored since the simplistic climate warming model requires conservative equilibrium which in reality does not exist.
Clouds are dissipative structures with the persistence of emergent attractors. All climatic features have spatio-temporal structure, such as ocean circulation, atmospheric circulation cells, anticyclones, depressions, storms, sea ice, etc. They are all dissipative structures. To understand their behaviour you need to model them as such, not in the incorrect conservative paradigm.
The behaviour of dissipative and conservative systems is fundamentally different. For instance, in conservative systems the Onsager reciprocal relations apply between pressure and temperature.
https://en.m.wikipedia.org/wiki/Onsager_reciprocal_relations
But this is not necessarily so in dissipative systems.
Noether’s theorem is important here.
https://en.m.wikipedia.org/wiki/Noether%27s_theorem
Her theorem is about physical systems having their own conservation laws. It is connected with the “principle of least action” which states that any system perturbed by a changing parameter will respond to this alteration by changing its overall state as little as possible.
Quoting the wiki on Noether:
All fine technical points aside, Noether’s theorem can be stated informally
If a system has a continuous symmetry property, then there are corresponding quantities whose values are conserved in time.[4]
A more sophisticated version of the theorem involving fields states that:
To every differentiable symmetry generated by local actions there corresponds a conserved current.
The “conserved current” is an important part of Noether’s theorem. I suspect that it applies to atmospheric thermodynamics, such that the flow of heat into and out of the atmosphere from space is the conserved current. Therefore changing CO2 will result in minimum system rearrangement with no change to the overall in-out flow of solar heat energy, which is the conserved current.
See also:
https://ptolemy2.wordpress.com/2020/02/09/the-principle-of-least-action-calls-into-question-atmosphere-warming-by-co2/
Climate modelling is like Bandersnatch, to progress you need to make a series of correct choices. Failing to recognise climate as a dissipative system means that you’re out of the game at the very first choice.
Phil, I wonder why climate modelers don’t operate from the same understanding.
You might be interested in Jerry Browning’s recent paper on climate models. He posted a summary at Judy Curry’s site.
Jerry is a Ph.D. in applied math, who spent his career on Bounded Derivative Theory applied to atmospheric flows. His paper is very mathematical.
But his point is that climate models apply incorrect equations of motion, that would cause the model to blow up did they not also include large dissipations, such as their hyperviscous atmosphere.
His paper proposes the fix. If you enjoy deep math, he’s where to go.
Thanks Pat
Jery Browning’s paper is a very important insight, it’s hilarious that the modellers need to make the air as “viscous as molasses” to stop the atmosphere from “blowing up”. They’re just numerically crunching over-simplistic equations and applying all manner of tweaks and fudges to get a credible looking result. The core maths is in dire need of fundamental change.
The comment about turbulence is important (this is of course suppressed by their unphysical work-around of making the atmosphere viscous with too much dissipation). I’ve always wondered for instance if the emission height should be considered as smooth. What if it’s complex and chaotic with a 3D fractal surface? That might change heat flows. I remember many years ago seeing a paper where they modelled a supernova and found that the models stubbornly failed to work until the modelled the shock wave propagation from inside in a chaotic turbulent way instead of liniear. Then it worked. But I’ve no idea what paper that was.
Agree Phil. These modellers seem to be locked myopically into into their own complexities. They are behaving a bit like molasses themselves. Time they got out of the box.
Pretty much all research into climate physics ceased after the modelers took over, Phil.
Apart from the Lindzen/Choi Iris Effect hypothesis, I know of no other attempt to improve the physical theory of climate, since 1990.
The modelers have effectively destroyed the field, driving out the bona fide scientists. Tim Ball has a lot to say about that.
A sad day. They’ve cornered the limelight and the funding.
BTW the modellers’ excessive and blunt use of dissipation as viscosity does not I think mean that they are recognising the system as dissipative – in the Ilya Prigogine sense. Instead they are forcing a conservative system to dissipate and that’s where the problems are coming from.
Phil,
I will repeat my post from above.
The only way any results from a climate model could be trusted are:
The continuum errors in both the dynamical and physical equations approximated by the model are smaller than the truncation errors of an accurate (almost convergent) numerical solution.
Now let us discuss each of these requirenents in detail.
1. All current global climate models are approximating the wrong dynamical system (the hydrostatic system) of equations. This has been mathematically proved in my peer reviewed manuscript that appears in the September issue of the journal Dynamics of Atmospheres and Oceans and in another thread on this site.
2. The physical equations are approximated by discontinuous parameterizations that have large continuum errors and that violate the necessary requirements that the continuum solution be expandable into a Tayor series. The necessary unrealistically large
dissipation needed to prevent the model from blowing up due to these discontinuities leads to a large continuum error and destroys the numerical accuracy as shown by the Browning, Hack, and Swarztrauber reference cited in the above manuscript.
3. As the requirements for a numerical method to converge to an accurate approximation of the continuum equations are violated,
the numerical solution will never be close to the true solution.
Thanks!
Governments are busy legislating life changing iniatives to tackle a problem projected by models which are incapable of projecting anything meaningful.
Meanwhile, it is becoming obvious that the climate of our water planet is contolled by water in one form or another. It impinges on the energy balance by interaction with short wave radiation and long wave radiation. It is involved in albedo via clouds, water, snow and indirectly ground albedo via rain and flora. It has a major contribution in energy transfer via ocean currents, winds and convection at the macro level and through its massive latent energy of phase change at the molecular level.
Yet carbon dioxide has been almost the sole focus of climate science for decades. While the alarmism has grown massively through hype, the understanding has not advanced at all. The science becomes less settled by the day. The scientists don’t like to discuss water and its clouds. They do not understand it and cannot model it. They do not recognise that this fact renders their projections useless.
Clouds are the elephant in the room. move along, nothing to see here.
Thoroughly agree. I think this goes back to the inauguration of the IPCC which was given the SPECIFIC remit to assess the risks of anthropological CO2 emissions which in essence removed the need to deal with the climate as a whole.
It is not surprising therefore that risks were found; as otherwise the IPCC would have been disbanded.
To me the most grievous ERROR perpetrated by the IPCC was the statement that water provided a POSITIVE feedback to its calculations of the GHE. This done by deliberately or otherwise by ignoring the thermodynamics of water, particularly at its evaporative phase change which occurs at a Planck sensitivity coefficient of zero.
This last statement opens up a large area of debate and discussion; but sadly, due to the politicisation now apparent this has been well suppressed as considered part of the sceptical camp.
The principle that for every force or influence an opposing force or influence is generated somehow appears to have been lost amid the complexities.
Good post. Thanks to the commenters for the interesting discussions within the comments.
My request: How about some of you engineers and physicists well learned in thermodynamics and phase change please weigh in. Be detailed and thorough and explain in terms for those that may be less knowledgeable. Phase change seems to be needing a helping hand in understanding clouds. The details in the cloud issue is ready to be solved.
I’ll have a go eyesonu; but my computer presentation techniques are very substandard and to date I can’t put the images I have into these comment sections. They are .ipg images and I don’t have the url. Can anyone help? Just hope the solution is simple as I have fat fingers.
Regards
Alasdair
As mentioned earlier, there is an excellent post here:
https://notrickszone.com/2020/09/11/austrian-analyst-things-with-greenhouse-effect-ghe-arent-adding-up-something-totally-wrong/
The author uses real NOAA airport cloud observation reports and finds that cloud cover is associated with higher temperatures, suggesting strong long wave warming. The is not what the conventional wisdom (based on models) expects and the finding undermines the whole importance of carbon dioxide as an agent for warming.
The paper discusses the role of clouds in warming and cooling and I found it informative and entertaining.
SC
Yes there are a lot of really interesting implications of Erich Schaffer’s paper.
It’s counter-intuitive that clouds warm, we all experience that they cool us on a sunny day.
At night of course they trap IR – but even in daytime they also downradiate IR.
Clouds are in the end just another form of energy.
Solar energy evaporates water which makes clouds.
Miskolczi recognised this and suggested looking not at atmospheric temperature but air energy content including temperature and water content, involving Virial theory.
For some reason he was of course violently pilloried for this – from a mixture of racism and political chicanery – defending a lucrative catastrophist narrative.
Schaffer’s discovery that clouds exert net warning changes our understanding of ENSO in the equatorial Pacific. Generally the east equatorial Pacific is cloudy during el Nino and clear during La Nina. This has been interpreted that the sea loses energy during el Nino and gains it during La Nina. But it could be the opposite.
“… Miskolczi recognised this and suggested looking not at atmospheric temperature but air energy content including temperature and water content….”
Once that energy is captured and contained within the water vapor in the air it is only temporary as that parcel of air will sooner or later convect upwards to radiate above the emission level. It’s the energy storage mechanism for the formation of the thunder heads or major storms. It’s just a temporary pause but would possibly skew the temperature records.
I agree that the water vapor energy (in its vapor state) needs to be included in any temperature records or they are not fit for purpose.
You seem to be talking about enthalpy. All the climate models try to use temperature as a proxy for enthalpy but it is a poor, poor proxy. You’ll never get a complete picture of the physics with temp as a proxy.
Too true Tim.
At evaporation water increases it’s enthalpy by 694 Watthrs/Kg; but the temperature remains constant, for example.
Consideration needs to be considered as to what put the clouds overhead. Were they brought in from an air mass that was warmer and had a higher moisture content from another region? Also the very cold waters in the Aleutian Islands, it would seem to be reasonable that warmer clouds aloft with a corresponding lower air mass would cause warmer ground temps.
To your last point – yes.
However the Aleutians were chosen as a place where the climate models were showing the highest negative correlation of clouds and temperature – a cloud cooling effect. And the data showed the reverse.
“While we do know that clouds will likely amplify global warming…”
LMFAO.
Just exactly how do they “KNOW” that?! Envelope to the head ala Johnny Carson?!
Or perhaps the same way they “KNOW” that rising atmospheric CO2 will “drive” the Earth’s temperature, something it has never been empirically demonstrated to do.
Only in so-called “climate science” are ASSUMPTIONS treated as if they were FACTS.