Technique could help climate models sweat the small stuff
PROVIDENCE, R.I. [Brown University] — A team of physicists and mathematicians has come up with a statistical technique that puts the fine details back into computer simulations of large-scale phenomena like air circulation in the atmosphere and currents in the ocean.
Computer models are generally good at capturing the big picture, but they are often forced to ignore things that happen at small scales. For example, models of a planet’s atmosphere capture the large-scale dynamics of jets and airflows, but they don’t include small-scale dynamics created by things like clouds and localized turbulence, despite the fact that those dynamics can often influence the larger scales.
“There are simply too many numbers for the computer to simulate it at a reasonable speed,” said Brad Marston, a Brown University physicist. “It might take years to simulate a day of the atmosphere, which wouldn’t be good.”
The traditional approach to dealing with the problem is to simply lop the small scales off of the simulation. A few ad hoc ways of putting some of that information back in exist, but they tend not to be mathematically rigorous.
“These schemes have always suffered from the criticism that they lack predictive power,” Marston said. “You have to make a lot of decisions that you really shouldn’t have to make but you’re forced to make.”
In a paper published in the journal Physical Review Letters, Marston and his colleagues show a method of averaging out those small-scale dynamics in a way that is computationally tractable, which allows those dynamics to be simulated and their effects to be captured in a rigorous way.
“We’re retaining the degrees of freedom at the small scale, but treating them in a different way,” Marston said. “We don’t have to simulate all the little swirls, so to speak. We treat them by using their averages and the sizes of their fluctuations. It allows us to capture the contributions of these small-scale dynamics that would normally not be included.”
In their paper, the researchers used the technique to model air jets forming on a round surface. They showed that the method produces results similar to brute-force numerical simulations of the same jets.
There have been prior attempts to treat small-scale disturbances statistically, Marston said, but those haven’t fared very well. Prior attempts have treated disturbances as being homogeneous and assumed they were not traveling in any one particular direction.
“But that almost never happens in nature,” Marston said. “Turbulence almost always has some directionality to it. That directionality is what makes these kinds of approximations work. It makes these approximations tenable.”
The researchers hope that the method might make for more accurate simulations of a wide variety of natural phenomena, from how the churning interiors of planets create magnetic fields to how air flows across the surfaces of cars or airplanes.
The method could be particularly useful in modeling Earth’s changing climate because the technique can more rigorously capture the influence of cloud formation.
“Cloud formation is seen as the largest source of uncertainty in climate models right now,” Marston said. “There are famous examples where different climate models that have different ways of dealing with the clouds give you qualitatively different results. In a warming world, one model might produce more clouds and another might produce fewer.”
By averaging those cloud dynamics and then simulating them in the models, it might be possible to reduce some of that uncertainty, Marston said.
The team has already started working to incorporate the method in climate simulations, as well as simulations of ocean currents and problems in astrophysics dealing with the behavior of plasmas.
“There are a whole bunch of problems out there where we feel this could be helpful,” Marston said.
###
Generalized Quasilinear Approximation: Application to Zonal Jets
J. B. Marston, G. P. Chini, and S. M. Tobias
Phys. Rev. Lett. 116, 214501 – Published 27 May 2016
Paper: http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.116.214501
ABSTRACT
Quasilinear theory is often utilized to approximate the dynamics of fluids exhibiting significant interactions between mean flows and eddies. We present a generalization of quasilinear theory to include dynamic mode interactions on the large scales. This generalized quasilinear (GQL) approximation is achieved by separating the state variables into large and small zonal scales via a spectral filter rather than by a decomposition into a formal mean and fluctuations. Nonlinear interactions involving only small zonal scales are then removed. The approximation is conservative and allows for scattering of energy between small-scale modes via the large scale (through nonlocal spectral interactions). We evaluate GQL for the paradigmatic problems of the driving of large-scale jets on a spherical surface and on the beta plane and show that it is accurate even for a small number of large-scale modes. As GQL is formally linear in the small zonal scales, it allows for the closure of the system and can be utilized in direct statistical simulation schemes that have proved an attractive alternative to direct numerical simulation for many geophysical and astrophysical problems
Yep, sounds “settled” to me.
Well the thing about clouds, is that they affect what goes on underneath them, so you have to be making measurements underneath the clouds, not on top of them.
And I’m sorry to inform Brown University, but so far we have not seen any models correctly report on the large scale facts of what we know already happened. All 57 models give results different from the recorded history.
G
So …. in fact they DONT make a generally good job of capturing the big picture, because they get it wrong.
Duh.
george e. smith
June 3, 2016 at 12:26 pm
Well the thing about clouds, is that they affect what goes on underneath them, so you have to be making measurements underneath the clouds, not on top of them.
I’ve seen the reverse. A cold mountain sticking up into a moist, warmer airflow causing cloud to form which persisted downwind of the mountain.
Everything is possible in cloud formation.
It is precisely small scale non linear interactions which define thunderstorms which determine tropical climate, which is the fundamental weakness of current models.
General circulation models have no predictive value.
This new step will not change that.
Instead of wasting their time “revising” those FUBAR Climate Modeling computer programs …… those brilliant programmers should direct their talents and energy too performing a much simpler task of creating a Birth Baby computer modeling program whereby the pregnant female provides a “sample” of both hers and her baby’s biological father’s DNA ……. so that the BB program can create a visual “model” of all the physical attributes that her birthed child will possess, including a summary “list” of INHERITED mental attributes and health issues.
I guess my English is a lot worse than I thought. Let me try saying it again ; this time in another language.
a) The sun; being the source of SOLAR ENERGY, is above the clouds. Excuse me, that’s WAY above the clouds.
b) The clouds block SOME of that solar energy, when it finally gets to those clouds.
c) Some of that cloud blocked solar energy gets scatter reflected (actually refracted) back to space; and some of it gets absorbed by the clouds; actually by the water in the clouds.
d) The solar energy that is not blocked by the clouds or scattered as blue sky, ends up underneath the clouds which did not block that energy.
e) The solar energy that ends up underneath the clouds which failed to block it, lands on the ground, most of which is wet.
f) The solar energy that lands on the wet ground, called ocean, goes deep down and eventually is largely wasted as “heat” energy (noun), but some does biological things instead.
g) Since the solar energy that gets by the clouds and lands on the ground; wet or dry, is what “heats” (verb) the earth, there is some evidence that this is what we should measure to find out if it is getting hotter or colder.
h) Such measurements of the solar energy that lands on the wet or dry ground, are best made with sensors on the ground, UNDERNEATH the clouds.
i) Measurements made of solar energy above the clouds, such as by satellites have consistently given higher values that clearly are wrong, and exaggerate the amount of “heating” (verb) that solar energy under the clouds, that lands on earth can cause.
j) Ergo, the solar energy that gets underneath the clouds and lands on wet or dry ground, is best measured on the ground, underneath the clouds, where the real values can be obtained.
G
Otherwise, what I said before, pretty much says the same thing, but in that other hard to understand dialect.
[Inelegant but very effective phrasing. .mod]
George
You might like this, these are graphs of the daily rate of temp change at the surface stations divided by the solar forcing in watts for that station (calculated for each location). These are are then averaged into 10 degree latitude bands. showing temperature change in F/WattM^2
https://micro6500blog.wordpress.com/2016/05/18/measuring-surface-climate-sensitivity/
I’m sorry, but if a minuscule amount of CO2 is a big picture…
…then clouds are an even bigger picture
Hallelujah! When the history of this asinine scare mongering fiasco is written it will be shown that water vapour is the WHOLE story! It is the great “balancer”, with surface warming causing more evaporation and the resultant vapour carrying massive quantities of heat aloft to be rejected closer to it’s ultimate destination and condensing into high albedo clouds before returning to earth to pick up another load. If the planet tries to get hotter, this ” engine ” speeds up and produces more cloud.
Way to go, John. Now they will have to open more unemployment offices.
Sorry to be off topic, but the readership should know of an excellent takedown of climate models, using the poor performance of economic models to predict the future. Here’s the link–http://www.bloomberg.com/view/articles/2016-06-01/global-warming-alarmists-you-re-doing-it-wrong
To get a hot link to work, put it on a line all by itself and WordPress automagically takes care of the rest.
http://www.bloomberg.com/view/articles/2016-06-01/global-warming-alarmists-you-re-doing-it-wrong
“automagically” ….my new favorite word of the month !! Thanks…
That works this week. Next week, they may improve it so it won’t work.
jorgekafkazar: Such an optimist!
In the meantime, there’s always the “copy the link and paste it” move.
Thanks, great article! R.
If these techniques are truly able to make the climate models more accurate, then more power to them.
I fear though that they will add these techniques in such a way as to make sure that the results remain the same.
Then the modelers will proclaim that since they now “include” clouds, that nobody has any reason to criticize the outputs.
The climatologists should be very afraid that physicists are sniffing around their turf. That is the start of making this great stinking pile into a real science. Much more rigorous minds are coming to bear. Like the great eye of Sauron that sees the weakness in things. Just wait til Dad gets home!
@MarkW:
“Any technology, sufficiently advanced, is indistinguishable from bullshit”
Well you have to make the climate models accurate, before you can make them MORE accurate.
So far, they are not even approximately accurate.
G
Water vapor has so much more effect than CO2 this lack of accounting for clouds makes it entirely obvious why the climate models do not work.
It has never ceased to amaze me how eloquently these “researchers” can describe something they can’t do, and make you believe they did it.
“By averaging those cloud dynamics and then simulating them in the models, it might be possible to reduce some of that uncertainty, Marston said.”
Please include the error bars. And, also explain how one can average a dynamic.
“In a warming world, one model might produce more clouds and another might produce fewer.
By averaging those cloud dynamics and then simulating them in the models, it might be possible to reduce some of that uncertainty, Marston said.”
So they are averaging the right and the wrong to get a wrongish but more certain answer? Have I understood correctly?
They propose averaging two wrong answers in the belief that the two wrong answers are normally distributed about the true mean. That is of course only true if the model is fundamentally correct in the first place. Prove it…
If it is fundamentally biased to one side of reality or the other then there isn’t much to gain by averaging the biased answers unless you have nothing else to do with your free time. It is as dumb as producing a ‘model mean’ using a bunch of models you know perfectly well are wildly wrong, all to the ‘high’ side. It shows true disingenuousness.
“””””…… That is of course only true if the model is fundamentally correct in the first place. …..”””””
There’s no evidence to support that. The symmetrical Gaussian distribution is the exception rather than the norm.
Most real distributions in real physical situations are not symmetrical. The Maxwell Boltzmann distribution that crops up all over thermo-dynamics for example is quite asymmetrical.
Our Temperature scale is bounded on one end by zero kelvin, but completely unbounded at the other end.
So it is highly unlikely that any physical process involving Temperature would ever yield results that would be symmetrically disposed about some ” proper ” value. Gaussian distributions are more fictional than real.
G
Well if you average the cloud dynamics, it is absolutely certain to be not accurate.
The MEASURED values are what is accurate. Whereas the AVERAGE can even give you a values that was NEVER measured by anybody, anywhere at any time; so it is completely fake and fraudulent.
G
“In their paper, the researchers used the technique to model air jets forming on a round surface. They showed that the method produces results similar to brute-force numerical simulations of the same jets.”
On the scale of the real world, convection cells start small, only a few meters across (dust devils) to a few tens of meters across (F1 tornadoes) to several kilometers across (thunderstorms and F5 tornadoes). Even at the scale of an F5, the ‘surface’ is flat. Their model may apply to large cyclonic storms, but most are not so large.
Until they can model the small scale instabilities that produce updrafts and downdrafts from which the Coriolis effect creates turbulent circulation and vortices, they are ‘not ready-for-prime-time.’
On the scale that we are talking about, the coriolis effect has no impact.
Coriolis seems to work on my kitchen sink – just how small a scale are we talking about here?
What do *you* think induces cyclonic circulation in hurricanes, typhoons, and tornadoes, if not the Coriolis effect?
What is the boundary line between Coriolis scales, and non-Coriolis scales, and what sort of thing happens at that boundary ??
G
One should not be giving all the credit to the “Coriolis effect” until such time that it has been proven that the action of “fluid logics” [gas or liquid fluidicis] is not the primary driver of in-line directional flow and/or “rotational and vortices” movement in minor to major volumes of gasses or liquids.
Me thinks a prime example of the action of “fluid logics” is the ….. Thermohaline circulation (THC) or the “great ocean conveyor belt”, to wit:
http://stempreacademy.hawaii.edu/sites/default/files/c-more/ocean-conveyor-belt/ocean-conveyor-belt-light_375px.jpg
Read more about fluid logics @ur momisugly https://en.wikipedia.org/wiki/Fluidics
Why exactly would any “Thermo-haline” gradient prefer one direction of circulation over the opposite direction ??
The rotation of the earth has had the same direction of rotation for as long as we have been observing it, and nothing more than that, and the arrangement of the land and ocean masses is required to explain the gulf stream.
So the surface waters along the East coast of South America are not flowing anywhere, in either direction. There’s no Japan Current taking warm water North into the Arctic, and sending cold surface waters south along the west coast of North America. WE have cold surface water coming down from Alaska, NOT warm surface water.
The “thermo-haline” theory comes somewhat later in the credibility index, than either string theory or parallel multiverses. (strictly MY opinion of course; not for general dissemination, or PhD Oral Presentations).
PS I do NOT dispute published ocean current observed directions of motion; just what drives them.
G
george e. smith – June 6, 2016 at 9:02 am
george e.,
Given my limited experience with “fluid logics” I believe you have unknowingly asked the wrong question, …… with the correct question being: just what pulls them?
In other words, me thinks that when “molecular segment Y” of a gas/liquid current is forced into motion ….. then that “segment Y” does NOT force, push or drive “segment Z” in front of it forward, ….. but on the contrary, ……. “segment Y” pulls or sucks “segment X” along behind it.
Given the above sucking (cohesion???) motion then a “Thermo-haline” gradient would have no preference for one direction of circulation over the opposite direction.
I think the really important question is: Just how in ell did the THC “current” initially get started?
Did the THC “current” form as a very small “loop” many eons ago when Pangaea was the only continent …… and then grew larger and convoluted as tectonic processes created the current continental configuration?
So many questions, so few answers.
Well a famous Nobel Prize winning Physicist told me, over a beer, sitting out in a colleague’s cool garden, that no matter HOW SMALL a section of a sphere you select, the curvature of that piece remains exactly the same. It NEVER EVER goes flat.
G
I wanted to ask him about blackbody radiation (fictional), and also quarks (he knows quarks), but I was too embarrassed, so we talked about dinosaurs instead. (He had been out dinosaur hunting) He works(ed) at SLAC
Now, about model validation……
That sounds like Step 1.
Actually, step one is the observation of the phenomenon. That comes long before one tries to model it.
Give me the top ten Physical theories that were simply dreamed up in somebody’s head before anybody had observed any phenomenon.
Yes there are breakthrough theories, of unexpected phenomena, that subsequently are observed experimentally.
But those theories are ALWAYS developed as a result of the FAILURE of some already well established models and theories, of well known phenomena, to correctly explain some observations.
But I’m eager to learn about those top ten “gee whizzes”
G
“…puts the fine details back into computer simulations of large-scale phenomena like air circulation…” back into? No, they were/are not in the simulations.
“but they don’t include small-scale dynamics created by things like clouds and localized turbulence, despite the fact that those dynamics can often influence the larger scales…” Yep, clouds are small scale but “influence” the larger scales. So they aren’t important for an accurate model but they are important for an accurate model I guess.
“A few ad hoc ways of putting some of that information back in exist, but they tend not to be mathematically rigorous.” ad hoc you say? So somebody is just pulling numbers out of the air to come up with an answer that someone desires? And if the models contain “ad hoc” information, what does that say about the results of the models?
“Cloud formation is seen as the largest source of uncertainty in climate models right now,” Marston said. “There are famous examples where different climate models that have different ways of dealing with the clouds give you qualitatively different results. ” So there’s uncertainty in the models? From more than one source? Is this what you mean by settled?
Brown has got to stop letting those unemployable English majors write the press releases without vetting them with the professors.
So “might” might reduce margin of “might”?
😀
Might makes Right !
g
Even with additional complexity, chaos always wins in the end.
Buzzkill.
Let’s see what happens when you tweak garbage. Maybe you come up with perfume. Dunno.
or enhanced garbage.
Or Twerked Garbage
..Twittered Cow Droppings ?
enhanced garbage is valuable for some folks.
Paywalled, of course. I wonder if Brown is attempting to patent the approach, or maybe the code.
Just typical academic journal behavior. It looks very mathematical so unless you have a PhD in applied mathematics you probably wouldn’t get much out of it.
The abstract reminds me of the lyrics from the chorus of Both Sides Now by Judy Collins:
“I’ve looked at clouds from both sides now
From up and down and still somehow
It’s cloud’s illusions I recall
I really don’t know clouds at all”
They don’t know clouds at all.
Joni Mitchell. Comes up every time clouds are mentioned on WUWT.
“Cloud formation is seen as the largest source of uncertainty in climate models right now,” Marston said.
It’s a good thing Brad Marston doesn’t live near California. He might be tried as a blasphemer for even suggesting that climate science isn’t 100 percent and irrevocably settled. There is no uncertainty in climate science according to California politicians. And they should know because they are the world’s leading scientists and the smartest people around. Just ask them.
Clouds are not ‘small details’. They are generated by small scale mechanics, but they are large scale phenomenon. I am highly skeptical of a statistical system purporting to predict cloud behavior.
Statistics is manipulation of what we already know (exactly). It is not predictive of anything we do not know already. The act of prediction is a leap of faith.
G
More fudge to go around.
As usual they offer no verifiable tests to judge whether or not this tweak provides realistic results.
But I am sure there will be numerous publications using this technique and confirming that the pet theory was not influenced by clouds, water or other inconvenient non-linear factors.
It is my understanding that the GCMs are limited by our current computing power to cell sizes of about 100 km. Until the resolution can be improved by 3 to 4 orders of magnitude (to 10 – 100 m), they fail to capture the relevant detail and interactions of these complex self-organizing dissipative systems (atmosphere, cryosphere, hydrosphere, lithosphere, heliosphere). These quasilinear approximations may help cheat that a bit, but you still need correct parameterization of your forcing factors. There is still a long way to go before we can build a decent model of something as vast and complex as earth’s climate. That doesn’t mean we should stop trying, but don’t bet the farm on the results of our current crude attempts. Hell, I wouldn’t even bet on one of Toneb’s 3-day forecasts.
Design is the problem, not resolution.
Truth!
Actually you are talking about a sampled data system. So compliance with the Nyquist sampling theorem, is the very first thing you MUST achieve or your “measurements” are not even valid data.
And (b) your model, if correct, should agree with what is observed in the real universe.
So that means your calculated model values should all agree with the already known measured values at those points in the real system where you have actually made real observational measurements.
For example, if you are designing the Electron Optics of say a Cathode Ray tube; the “gun” and the deflection optics, using an electrolytic tank (one method), you can probe around in a two dimensional section of the three dimensional system, with a voltmeter probe, and plot iso-potential lines giving a map of the electric fields.
Then you can use Finite Element Analysis (FEA) to create expected values at points not yet measured, to look for parasitic effects that might be in need of corrective design alterations.
It is quite pointless to calculate a map of supposed real physical values of variables, and not be able to check them against real observed values at known locations.
G
Cmon people, why be so negative about a subject you clearly know little about? Treating the relations between mean flow and fluctuations goes back to Osborne Reynolds and after a century of development those methods have about run their course. These guys are advertising a novel approach based upon a spectral decomposition (perhaps capturing something like “large eddy simulation” method results at lower cost?). If the small scale interactions are locally statistically related to computable scales in a reproducible way (a testable assumption) this could produce worthwhile improvements in GCM fidelity with only modest additional computational cost. Will it work in climate simulation context? Time will tell, off the cuff criticism won’t.
CarlH June 3, 2016 at 3:34 pm
Thanks, Carl, but c’mon, why be so positive about a subject you clearly know little about? At present, global temperature results from the current crop of climate models can be very closely approximated (R^2 over 0.90) by merely lagging and scaling the input forcings. They are mere mechanical automatons, varying only in their inputs.
Not only that, but nobody has ever shown that the fluid equations used in the models actually converge. It’s one of the reasons that they don’t use finer gridcells—it gets unstable quite rapidly.
Add to that the fact that they cannot replicate the historical drops in global temperatures (e.g. 1940’s to 1970’s) without bogus specially-chosen values for their aerosol forcing … and often can’t replicate the drops at all. Take a look at the individual CMIP5 runs, many of them are a joke.
And this doesn’t even start to mention that the models have not been subjected to V&V (Verification and Validation) even as strong as the V&V applied to elevator software …
Nor have I ever seen an analysis of a “failed” run, or even a report of a “failed” run, so you can bet that all of their runs in the literature and in the archives are specially selected. We never get to see what gets left on the cutting room floor.
So yes, it’s possible that this new technique may assist as you say … but until some other major problems are solved, I fear it may well be lipstick on a pig.
Finally, I note that the majority of the negative comments are not about the new method itself. They are about the foolish, and in some cases even incoherent, claims made by the press release.
Best regards,
w.
PS—For more info on how mechanistic climate models actually are, and how their output is just lagged and scaled input, see:
You should enter this in the Bullwer Lytton prize competition. You’ve got more three dollar words in less space, than was previously thought possible.
Some of us make no bones about not knowing a whole lot about the climate in places we have never been. What we are concerned about is the claims made by others about what they think they know, about stuff they have never even measured.
But thanks a whole bunch CarlH; all my life I have wondered who the hell it was that came up with the Reynolds Number.
G
Pick a cell . . . any cell. Where do its initial values come from?
I pick 200 miles west of Easter Island. Got data?
“All previous models were junk. Our new, improved one isn’t so good. It will be better in the future.”
This will simply mean a rewrite of the cloud mechanics followed by tweaking of the model’s parameters to (allow the modelers to) get a better historical fit and we’ll have models that still can’t accurately project but we’ll not know it for another decade or two.
Did some research before commenting. Their new method, while interesting, does nothing to solve the fundamental climate model problem of attribution in parameterization. They still parameterize via ‘spectral averaging’. Clouds are still parameterized, only differently. So how much of the ‘new’ cloud parametization is anthropogenic, and how much is natural variation?
Their ‘proof’ of their method’s validity on atmospheric jets from a spinning sphere does NOT address or resolve the climate attribution problem in any way, shape, or form. Logic fail.
Quasilinear: Now there’s a word that screams pseudo science.
But will it impress the girls?
Democratic girls.
Actually I have just about enough basic math to understand what they are saying a little bit, and I think you may be misjudging these people.
If I understand correctly, this could actually be a fundamental breakthrough in non linear system analysis.
And if it is, and it is accurate enough, and it gets put into all those GCMs then they WILL start to predict at least more accurately, and guess what. That could be the end of AGW as we know it.
And its the perfect let out for climate scientists ‘well the models weren’t good enough, and that’s why we raised the alarm’ .
I think we should watch this one. It has a slight tremor of authenticity about it.
Rem,ember these are not climate scientists here. These are mathematicians. And physicists.
If nothing else, it might lead to a CFD package worth billions.
It’s a five sillabel word for nonlinear which is only four sillabels, just in case you are trying to write a Haiku.
g
ps. Nowhere in my entire thought process, about the role of clouds in earth climate, does it EVER occur to me that some of those clouds might be moving around microscopically and macroscopically, locally and globally, so that one needs to check on Reynolds numbers to determine the relevant fluid flow processes that might be going on.
To me, clouds are 4H-club dirt simple.
Solar energy runs into clouds; whatever clouds there are. Some of the solar energy is refractively scattered back out into space. Some additional components of the solar energy gets absorbed in those clouds; whatever clouds there are.
None of that solar energy ever makes it down to the surface of the earth as solar spectrum wavelength EM radiant energy, and get mostly absorbed in the ocean, with minor reflection from the ocean, or absorbed in the solid surface, with somewhat larger reflection from the solid surface.
The only thing that is of any consequence in those interactions, is HOW MUCH incoming solar energy doesn’t make it to the surface absorption and reflection interface, and that depends on HOW MUCH cloud there is and where that cloud is relative to where the solar radiant energy is arriving.
The solar energy is traveling at the velocity of light; about a one millisecond transit time through the atmosphere, to the surface.
The most violent atmospheric particle motions ever observed or conjured up by play station, don’t even show up as having moved at all, and affecting the resultant photon trajectory to the surface.
G
[When misspelling sylables incorrectly, ’tis best to have many sillybells as possible .. .mod]
In climate, it matters how cloud behaviour changes over decades. Working out a better method of dealing with “averaging out those small-scale dynamics” adds little or nothing to actual knowledge of cloud behaviour. Even if there is improvement, when fed back into one of today’s “climate” models the improvement will show up for a very short time only. I doubt they could get a single day out of it. The reason is that the “climate” models are really only low-grade weather models that are incapable of predicting anything. They can never have any value for climate prediction while they retain the same basic structure and methodology. https://wattsupwiththat.com/2015/11/08/inside-the-climate-computer-models/
If these researchers do find out more about how clouds work, then good luck to them. But until they feed results into a climate model (not a weather model) they won’t get a climate forecast.
Aye, there’s the rub. When I found I had to substitute a statistical routine in a model where I’d intended to use a determinate one, I knew I had lost. The model was doomed to fail eventually once it left the bounds of the statistical validation period. Often sooner.
When the average happens it is too late to do anything about it.
Nothing in the entire universe that is NOT human homo sapiens sapiens, is even aware of or responds to an average value of anything, or even the average value of everything.
Real systems operate in real time, and do everything they can do immediately. They do not wait for anything before doing anything, and nothing happens before it can happen, nor does it wait until it feels like happening. When it is possible to happen, it happens, not before nor later.
G