Guest Post by Willis Eschenbach
PART THE FIRST – REAL WORLD
Well, my monkey mind started thinking about the relationship between how much sunshine is absorbed at the surface and the surface temperature. Here’s the CERES satellite data showing the relationship:
Figure 1. Surface temperature and how much solar radiation is absorbed at the surface. Radiation is in watts per square meter (W/m2).
As we’d expect given our daily experience, more sunshine raises the temperature and less sunshine lowers the temperature.
The question naturally arises—just how much does the surface temperature go up for each additional W/m2 of absorbed radiation at the surface?
We can approach this question in three different ways. First, here’s a scatterplot of the monthly values shown in Figure 1, along with the trendline.
Figure 2. Scatterplot, monthly surface temperature versus how much solar radiation is absorbed at the surface. Because there is uncertainty in the monthly averages, I have used Deming regression rather than standard linear regression.
The second way to calculate the relationship between the surface temperature and surface absorbed sunshine is by linear regression on a gridcell basis, weighted by the area of the gridcells. This gives us the same answer, 0.22 °C per each additional W/m2 of absorbed solar radiation.
The third way to examine the relationship looks at the long-term averages of both temperature and absorbed solar radiation as a scatterplot on a gridcell-by-gridcell basis. This allows us to see what is happening at different temperatures.
Figure 3. Scatterplot, gridcell averages, temperature versus average absorbed solar radiation. The slope of the red line shows the trend of temperature with respect to absorbed solar radiation at various temperatures.
There are several interesting things about this graph. First, over most of the earth (central region above), the relationship between absorbed solar and temperature is pretty linear, with an average trend (slope of the red line) being 0.21 °C per W/m2.
To assist in understanding this, here’s a graphic showing how much solar radiation is absorbed at the surface.
Figure 4. Surface absorbed solar radiation (downwelling radiation minus reflected radiation)
You can see the effect of the nearly continuous clouds at the Intertropical Convergence Zone (ITCZ) as a yellow stripe just above the equator.
To return to Figure 3, in the areas where there is little sunlight, the temperature increases very rapidly with increasing sunshine. Here is a map of where those areas are.
Figure 5. Parts of the world where the annual average absorbed solar radiation is less than fifty watts per square meter.
And at the right-hand end of the scale in Figure 3, surprisingly, in areas where average absorbed solar is above about 210 W/m2, increasing the absorbed sunlight doesn’t warm the surface much at all. The average response in the colored areas shown below is 0.03°C per W/m2. Go figure. Here are those locations.
Figure 6. Parts of the world where the annual average absorbed solar radiation is greater than two hundred ten watts per square meter.
The horizontal dotted lines above and below the Equator on the map in Figure 6 show the limits of the tropics. Note that most of the tropical oceans don’t warm much further in response to absorbed solar radiation increasing byond 210 W/m2.
And at the end of that, we have three different estimates of how much temperatures go up when solar goes up and go down when solar goes down. All three of them are on the order of a 0.2 °C temperature change for each 1 W/m2 change in absorbed sunshine. And all three show that temperature varies with absorbed sunshine, going up with more sunshine and down with less sunshine, just as we see every day.
PART THE SECOND – MODELWORLD
After I looked at what is actually happening, I thought I’d take a look at what the models say is happening. The model data is available at the marvelous KNMI website by selecting “Monthly CMIP5 scenario runs. These are from the Fifth Computer Model Intercomparison Project (CMIP5). The surface air temperature is identified as “TAS” (temperature air surface). Downwelling solar at the surface is “RSDS” (radiation shortwave downwelling surface), and reflected surface solar is “RSUS (radiation shortwave upwelling surface). The absorbed radiation is the downwelling solar minus the reflected solar.
I started with the temperature data. I was interested in the historical data, which is essentially identical for the four “Scenarios”, yclept RCP26, RCP45, RCP60, and RCP85. I used the RCP26 data. The historical data ends in 2012. Here’s the CMIP5 mean global historical surface temperature reconstruction compared to the Berkeley Earth global surface temperature.
Figure 7. Berkeley Earth and CMIP5 model temperatures compared.
Well, that’s pretty respectable. The models have done a decent job of emulating the major changes in the historical temperature. (It does bring up the question of how different models with widely differing climate sensitivities can all hindcast the temperature so well, a question I discussed in “Dr. Kiehl’s Paradox” … but I digress.)
Having done all of that, I went to look at the modeled absorbed solar radiation … and my eyes bugged out of my head. Here’s that modeled result. As with temperature, the solar results are basically identical for the four scenarios, so I’m showing RCP26.
Figure 7. CMIP5 RCP26 historical surface absorbed solar radiation anomaly.
YIKES! Temperatures are going up and absorbed sunlight is going down? Say what? How unbelievable is that?
But wait, as they say on TV, there’s more! Here’s the same RCP26 solar data, this time including their projection of absorbed solar at the surface out to the year 2100.
Figure 8. CMIP5 RCP26 historical and projected surface absorbed solar radiation anomaly.
It’s … curious. Modeled surface absorbed solar is decreasing over the historical period all the way right up to 2012, and then it immediately turns around and starts increasing.
Probably just a coincidence.
But consider … if they somehow get rising historical temperatures with falling historical absorbed solar radiation, think of how high their future projections will be with rising absorbed solar radiation. It’s a win-win situation for the alarmists!
And how come I’m the guy who notices these things and not the good folks running the models or the people at CMIP5?
Always new questions.
Here on the northern California coastal hillside where I live, we are in the midst of an all-too-frequent occurrence … Pacific Gas and Electric, aka PG&E, can’t keep the electricity on. Once again, we’re out of power. Sigh. I just fired up my reliable fossil-fuel-powered Honda i2000 generator, strung the extension cords, and I’m back in business.
But never fear, the geniuses running California have a brilliant solution for the endless outages, brownouts, rolling blackouts, and power shortages.
You ready for their plan? Here it is:
- Increase even further the amount of unreliable, intermittent renewable wind and solar electric generation
- Jack the already outrageous ($0.34/kWh) electrical prices even higher to discourage demand
- Increase the grid load by banning the sale of gas-powered lawnmowers, leaf blowers, garden tractors, chainsaws, etc. by 2024
- Increase the grid load further by banning the sale of gas-powered generators like mine shown above by 2028
- Close the one remaining nuclear power plant, and
- Drive the load on the grid through the roof by forbidding the sale of gas-powered vehicles after 2035 …
Yeah, that’s the ticket. That’ll reduce the brownouts, rolling blackouts, and total outages. Plus it will screw the poor today, but hey, we’re helping the poor in the year 2050 and saving the world, so it’s all for the best in this best of all possible worlds …
Buncha fricken’ rocket surgeons, alright.
My best to all,
w.
You know it but I still gotta say it: When you comment, please quote the exact words you are referring to. This makes everything clear, and avoids the misunderstandings that seem to proliferate on the intarwebs.
That’s a good generator.
Clearly you meant 2035. Even Newsome can’t ban things in 1935.
Thanks, fixed. And if Newsom could he certainly would.
w.
“”..how come I’m the guy who notices these things and not the good folks running the models or the people at CMIP5”?
You are a voice in the wilderness; an heretic.
Yeah, but it’s their damn job, it’s what they are paid to do …
Grrr.
w.
Their job is to push a narrative. That demands a pushback – and someone intelligent enough to do it…
That’s a quaint notion.
As has been observed before “the only way to get pertinent answers is to ask impertinent questions” – unfortunately most of the climate enthusiasts are so enamoured of their own cleverness they ignore or cancel all impertinent questions.
Willis,
These discrepancies have long been noted by some others. When CMIP results first appeared before us, we wrote questions like “How can the models get the right result when some inputs are based on faulty data, like proxies calibrated to temperature using an adjusted temperature record?”
It resembles walking on a waterbed in pursuit of imaginative sex. You put your foot down in one spot, then the whole mattress surface changes shape. By analogy, you input some dummy miscalculated proxy temperature, the whole model surface of results changes. The force of the wrong input produces wrong outputs. What you are doing, Willis, is discovering and displaying beautifully, some of these distorted outputs. Something has to give and it shows.
Those lovely scatterplots with patterns in waves might have more information content yet to be discovered. When there is a repeating pattern, what is its frequency and what is its cause, that type of puzzle. What would a new plot look like if you added or subtracted (say) 1 W/m^2 from the input data behind the model in your figure 3 scatterplots? Would the temperature platform at 30deg C be forced higher, despite constraints to 30 C like you and Ric W have often shown us?
My apologies for asking more Q than A, but I am getting a bit old to calculate much any more.
Geoff S
Maybe they did notice. They may even have expected it and said “oh yeah, that again”.
Because Willis follows the data, not an agenda
considering about 1/3 solar absorbed is into atmosphere, and that nearly 50% of the incoming solar beam is thermal IR, increasing water vapor should increase atmospheric solar interception.
In terms of the non-vapor condensate i.e. clouds, they have an average emissivity of about 0.6, so it is presumable that cloud IR reflection is about 0.4 and absorptivity 0.6.
At the surface, the vapor pressure saturation curve is becoming almost vertical at the upper end of terrestrial temperatures. And so, surface partitioning of turbulent flux is increasingly favoring the evaporation at warmer temperatures, and increasingly favoring the sensible heat flux at lower temperatures. At about tropical oceanic temperature under direct sun nearly all heat flux is evaporation.
http://www.atmo.arizona.edu/students/courselinks/fall12/atmo336/lectures/sec1/satvap.jpg
It is almost all pointing to a negative feedback regime. The models might be seeming out of whack because they are calibrated to historical observation while simultaneously assuming a positive feedback regime. This would cause some strange behavior.
Willis, if I am reading the graphs correctly, the measured solar energy absorbed at the surface per Ceres is flat from 2000 to 2010 and then slowly increasing from 2010 to 2020. The values used by the model (figures 7 and 8) disagree significantly in both the value versus measured and the trend. If you are constructing a model and you chose to ignore real data that should be part of the model, you have to explain why you are not using those data in your model. Anything else is cherry-picking at best and complete dishonesty or worse. Any explanation from the authors or users of these models why they don’t use a significant input? To me it looks like tuning the model was more important than figuring out why the model needed tuning in the first place.
On your second thought, it is hot today in Texas. We are using 71.6 GW with 5.8 GW of committed spare capacity. Solar is providing 17.8% of our power while wind is providing 10.6% of our power. Natural gas is providing 50.7%. I pay about $0.11 per kwhr. I grew up in California but i could never afford to move back nor would I want to.
You can’t afford to move back to California, and I can’t afford to leave the place …
w.
Well, kinda. It actually goes up and back down again from 2000 to 2010, as does the temperature.
w.
What’s curious is the sun’s sinusoidal change in luminosity over about a decade has no apparent effect on the earth’s energy imbalance. It would appear that the Earth is controlling the energy imbalance quite well. I would say there is not much we can do about this but to adapt. And we won’t have to adapt to much.
I keep asking for government to tell us how many degrees per dollar spent does fighting climate change produce? No one will answer my question. I keep reminding them that it’s their plan, not mine. If they don’t know, then they can’t really plan at all, can they?
Doon, I once showed that if the US went to net zero tomorrow, by 2050 IF the IPCC is correct it would be about 0.1°C cooler in 2050. It would cost on the order of ten trillion dollars to do that. You do the math.
“Going To Zero“.
w.
Thanks. I remember that article. But I was talking about local governments, I should have made that clear.
Under 2006 California law AB32, all 58 counties must have a plan to reduce CO2 emissions to 1990 levels by 2030. Here in Santa Cruz County, Cemex closed all cement manufacturing in 2010 which immediately reduced the County’s CO2 emissions by 59%. I suggested then that all other county efforts were meaningless as we were already State compliant and 20 years earlier than required. So we did not need another plan, mission accomplished. The county disagreed. So I asked for the degree per dollar amount that additional, voluntary action would have. All I ever received were blank stares and crickets.
Funny story, Bonny Doon.
w.
Your problem was thinking that the title of the action was actually the purpose of the action.
Affordable Care Act
Inflation Reduction Act
If they drive enough people out of California maybe both CO2 emissions and electricity demand will fall enough to make it all good for them——oh, it just occurred to me: maybe they’re well aware of this and this is their true goal.
“Cemex closed all cement manufacturing in 2010”
So I presume absolutely nobody is using cement since then in your county?
’cause if anyone in your county imports cement from anywhere else, that would mean that your county’s good behavior was a wasted effort, with respect to saving the planet! And thus, the planet will be lost! OMG! 🙂
Worse than a wasted effort, because now you have to burn CO2 to get that cement back to where it used to be produced.
And what happened to the price of cement? I suppose they only make bitumen roads now? And concrete high rise is right out! So oil is back, how funny.
Foundations for windmills are made of?
more like 10 quadrillion
True, as what can’t happen, won’t.
Which model sounds more complicated, climate based on increasing concentration of a gas or global economy based on human reaction to the worlds biggest economy imploding? I would not bet on either.
I have long waited to learn at which point economists first correctly aligned their models with this thing termed the earth’s average temperature.
A curious mix of disciplines, I’m sure you will agree.
Same here. My (CA) city is requesting feedback on their draft Climate Action Plan. I asked:
1. How much will it cost?
2. How much will it improve the climate?
The response was… silence!
Simple truthful answers: lots and none.
“If they don’t know, then they can’t really plan at all, can they?”
Sorta like the Russian generals who think everything is going according to the plan.
Thats my “killer question” for when our politicians start campaigning here in the UK.
Gavin Newsom actually banned ALL California sales of gasoline engines under 25 hp by 2024. It’s going to be fun for Cal Fire to replace their chainsaws.
Can oil be used for blades?
Stihl’s largest chainsaw is the MS-881 Magnum.
The powerhead alone is 22 lbs. That is for a saw with only 8.6 bhp.
If the fire crews are only going to be allowed chainsaws with greater than 25 HP, then I predict that the burn areas of forest fires are going to greatly increase in California.
Of course, the journalists will attribute this increase to “climate change” – not increasingly stupid government policies!
In Woke-achusetts, the state is now giving only electric chainsaws to its park workers- which of course will only run about an hour per battery. One ranger put up a video saying he brings 2 batteries with him- ergo, he can work 2 hours per day.
On par with the average government worker’s day. Been there, watched that. When asked about how many people worked in my Division, I said; “Oh, about half of them.”
The problem is solved –
“Climate Change Turns Out To Be a 71-Year-Old Guy Named Ed
Climate change has been caught and is behind bars.”
https://www.frontpagemag.com/climate-change-turns-out-to-be-a-71-year-old-guy-named-ed/
Many small farm tractors are less than 25 hp. The LS 25 is 24.7 hp.
Data omission is science fraud.
And for the churnalists: A lie by omission is still a lie
Willis,
You asked, “It does bring up the question of how different models with widely differing climate sensitivities can all hindcast the temperature so well, …”
I thought that the answer to that was that all the models are tuned to the give good agreement to historical temperatures.
True, and that totally destroys the arguments that the models are physics-based and that the climate sensitivity is important. Clearly both claims are nonsense.
w.
I wonder how far we go with this argument? I ask because Newton’s law of universal gravitation has a free parameter that must be tuned to match observations. Does that mean it is not physics-based and/or nonsense? What about the standard model and its many free parameters that must be tuned to observations? Or what if we add up all the free parameters in all equations that describe reality? How many hundreds of them are there? Is science itself not physics-based and/or nonsense? Maybe free parameters are acceptable in all disciplines of science except that of climate?
The British rock band, Dire Straits, clearly understands the logic problem more clearly than you do.
“Two men say they’re Jesus, one of them must be wrong.”
Isn’t God supposed to be everywhere?
What about Queen? They’re also British and Brian May is a bona-fide astrophysicist. He’s way smarter than I’ll ever be.
‘I ask because Newton’s law of universal gravitation has a free parameter that must be tuned to match observations.’
Are you referring to ‘G’, the gravitational constant?
That’s the one.
https://www.nature.com/articles/514150a
The gravitational constant is just that, a CONSTANT. It is not a tunable parameter. Just because we can’t measure it without an uncertainty interval being associated it doesn’t mean it is not a constant. As usual with measurements we just don’t have precision enough in our instruments to eliminate the uncertainty.
You have *NEVER* understood uncertainty in measurements. As usual, you don’t understand it in this case. The uncertainty associated with our measurement of the gravitational constant *should* be propagated through any calculation made with it. If you just assume a value for it (i.e. make it a tunable parameter) then at some point you might just kill an astronaut trying to cut it close in orbiting some body.
Yikes! I’ll be brief because others have already weighed in on this. I think in comparing our uncertainty in knowing ‘G’ with the tuned slop in a GCM goes well beyond any other example of false equivalence I’ve ever seen. Oh, and at least Newton got the sign right.
I think you missed my point. My point is this. If tuning the free parameters of a model to match observations totally destroys an argument that it is physics-based then entire swaths of well established science including models of gravity, particle physics, etc. will need to be indicted of being non-physics-based.
BTW…Not only did Newton not even attempt to estimate G he thought it was impossible. The first direct estimate of G occurred more than 100 years after Newton hypothesized its existence. Think about that. Other than knowing that it must be positive he couldn’t even begin to estimate it. Yet no one dares to indict the law of universal gravitation as not being physics-based.
bdgwx writes “If tuning the free parameters of a model to match observations totally destroys an argument that it is physics-based then entire swaths of well established science including models of gravity, particle physics, etc. will need to be indicted of being non-physics-based.”
No you missed the point. G is a constant. The impact of clouds in a model is not. It’s not “physics based” to parameterise clouds. It’s a fit. And when you add that fit into a projection then the whole model becomes a fit.
We think G is constant only because experiments cannot falsify that hypothesis. Regardless being constant does not preclude it from being a free parameter. What would preclude it from being a free parameter is if it were defined exactly or it emerged from a model. It is not defined exactly nor does it emerge from any model. It must be estimated experimentally. Therefore it is a fundamental constant that is also a free parameter.
Newton’s universal law of gravitation is a fit as well. G has be tuned to match observations for it to be able to make predictions correctly. Einstein’s field equations is a fit too. Both G and Λ have to be tuned to match observations for it to be able to make predictions correctly. The standard model is also a fit. There are (currently) 19 free parameters that must be tuned to match observations for it to be able to make predictions correctly.
BTW…each of the models I mention above are fits not only because they utilize free parameters, but because their core mathematical structures are known to only be approximations of reality. So even if G, Λ, and the 19 parameters of the standard model were to somehow emerge from a deep physical principal or calculation or otherwise not require experimentation those models would still only be imprecise fits to reality. Yet no one seriously challenges whether they are physical or not.
As Tim tried to point out to you, G is a CONSTANT. The problem in defining it derives from measurement uncertainty. Part of that uncertainty is measurement precision. Part of it comes from using different devices as the graph shows. As we’ve tried to tell you with temperature measurements, taking many different measurements and claiming that you can increase the resolution through statistics just doesn’t work.
Do you really think that there haven’t been enough measurements taken over the years such that an average can be calculated, and the “precision” of the mean can be increased by dividing σ by the √n. Obviously, the resolution of measuring instruments have nothing to do with the actual value, since resolution can be easily increased through sampling statistics!
Remember, all you are doing is adding decimal points to a CONSTANT, and not dealing with a continuous temperature curve.
You say:
I can’t figure out whether you’re trolling, denser than a black hole, or just profoundly wrong. The first person to MEASURE, not “tune” or “fit” but MEASURE, G was Henry Cavendish.
He used Newton’s formula, viz:
F = G (m1 m2) / R2
He set up a very clever experiment whereby he could MEASURE m1, m2, R, and F. Substituting this into the formula let him CALCULATE G directly.
Note that nowhere in his experiment was there the slightest bit of “tuning”. And G is not a “free parameter”, it is EXPLICITLY SPECIFIED by the values of F, m1, m2, and R.
Give it up, my friend. You’re ignoring the First Rule of Holes.
w.
Use whatever ad-hominem you feel is best. I’ll even defend your right to do so. And you can take comfort in knowing that I will still treat you with the utmost respect. I think you’re a smart person; certainly smarter than I. I certainly don’t think you are dense as blackhole or otherwise. And I always learn from your posts.
I’m well aware of Cavendish. Anyway, I use the word “tune” synonymously with “via experimentation”. Because free parameters do not emerge from the model itself they have to be obtained using observations from one or more experiments. Once adequate values are obtained via experimentation for the free parameters then, and only then, can the model be used to its fullest extent.
It was clever indeed. Anyway, I can tune free parameters using calculation as well. I do it all of the time. In fact, I don’t know how one would do it without performing some kind of calculation even if only their head. The act of calculating is not what discriminates a free parameter from one that is not. What discriminates a free parameter is how it is determined. If it can only be determined via experimentation then it is a free parameter.
Except for the fact that G was determined via experimentation. And even then it wasn’t the correct value. In fact, nobody knows the correct value and we never will (barring a massive upheaval in our understanding of gravity anyway). All we can do right now is refine the value with further experimentation. Ya know…make adjustments to it…or tune it.
“All we can do right now is refine the value with further experimentation. Ya know…make adjustments to it…or tune it.”
You dont tune constants. That’s a misunderstanding of what tuning is.
Tuning cloud parameters alters other quantities in a model such as the TOA radiative imbalance and the cloud parameters are tuned so that the TOA radiative imbalance is close to what is observed. Tell me why that isn’t a fit in your view?
And I use “observed” very loosely. The TOA radiative imbalance is more “expected” given the measured values of ocean heat content changes, primarily.
And I use “measured” very loosely wrt ocean heat content as well.
“That’s a misunderstanding of what tuning is.:
He doesn’t understand because he doesn’t *want* to understand. He has never grasped the concept of measurement uncertainty because he doesn’t want to understand it. It’s why he and so many others in climate science believe the “best fit” analysis of a regression line is the uncertainty in the data (i.e. the standard error of the regression). It relieves them of the burden of actually having to analyze the data using the measurement uncertainty associated with the data. The meme “all measurement error is random, Gaussian, and cancels” is the unwritten assumption in everything they do!
Tuning is the process of making adjustments until a specific value is obtained. The specific value to be obtained in the case of G (a value we think is constant) is the one which makes observations most consistent. But there is no single value that makes observations maximally consistent. That’s the problem. We’ve been tuning G with more and more precise refinements for over 200 years with countless experiments and we still don’t know what its value is exactly and likely never will (barring a major upheaval in our understanding of gravity) due to experimental methodology and measurement uncertainty. And oddly today different teams get values that are not consistent with each other. For a “fundamental constant” that is a sad state of affairs. Maybe one day we’ll have a theory of gravity in which G emerges organically without the need for experimentation or better yet we discover there is a deeper underlying principal that negates the need for G altogether. But for now…tuning the value to make observations consistent is what we’re stuck with.
“The specific value to be obtained in the case of G (a value we think is constant) is the one which makes observations most consistent.”
Why do you INSIST on ignoring measurement uncertainty? What observations are you talking about being consistent? Both G and things like orbital influences are based on MEASUREMENTS, measurements that have uncertainty. If your orbital calculations are not consistent with observations is it because of the uncertainty of G or uncertainties associated with the measurements of the orbital path?
Again, the value of G is not an assumption that you are free to choose the value of. It is not a FREE parameter.
“Tuning is the process of making adjustments until a specific value is obtained.”
With constants this is just wrong. Constants are measured not tuned. And some constants are calculated from measurements using a formula. But they’re still based on measurements.
By comparison tuning is making choices so that other related quantities are closer to what is desired.
Can you not see the difference?
There is no physics based formula relating cloud parameters and the TOA radiative imbalance for example.
How many parameters from the Fluid Dynamics discipline are included in climate “science”?
good question for Mickey Mann- I’m sure he knows 🙂
after all, he knows how to measure tree rings- so he must fully understand Fluid Dynamics
Does anyone (particularly NOT Mann) have all the answers to the questions FD raises?
My layman’s perspective on this is that climate(s) behaviors are all about fluid dynamics.
As I think WE once observed (my paraphrasing)
– in climates, everything affects everything else, to varying degrees, constantly.
Cmon bdgwx, surely you can see the distinction between measuring g from observation, where there is only one parameter and only one possible value that can match observations vs 36 models each with multiple parameters which are all different yet purport to represent the same reality.
That’s Nitpick Nick level sophistry. You’re better than that.
I was actually talking about G; not g. Though g is also a free parameter used in various models as well. Neither were really the focus of my post though. G is but one among many examples of a free parameter.
While we’re on the topic of G I’ll provide some more commentary for people to ponder. I’ve often been told that because scientists cannot agree on some of these free parameters that necessarily means that climate models are useless, not fit for the purpose, not physics-based, etc. It turns out that scientists cannot agree on what G is either. Does that mean Newton’s law of universal gravitation or Einstein’s general relativity are useless, not fit for the purpose, not physics-based, etc. as well?
It is the size and effect of the disagreement that is at issue.
What is the threshold that discriminates between a physics based model and one that isn’t?
That’s not the question. The difference is between equations on the one hand and iterative climate models on the other hand.
And if you think the climate models are “physics-based”, please read (or perhaps re-read) my post “Meandering Through A Climate Muddle“.
Best regards,
w.
I don’t think invoking the iterative nature of model is going to help you in discriminating between physical and non-physical models. Afterall Newton’s law of universal gravitation must be applied iteratively to explain the motions of an n-body system. And even then it fails in explaining all of the observations. Does that mean Newton’s law of universal gravitation is not physical?
You are trying to compare the validity of concepts with the inability to EXACTLY measure a physical action. That is an ill posed problem to solve. Your argument fails.
“What is the threshold that discriminates between a physics based model and one that isn’t?”
F=ma is a physics based model because the relationship holds irrespective of choices of F, m or a.
Climate from a model is not physics based because the calculation includes clouds which are parameterised and those parameters are tuned so that the TOA radiative imbalance is as expected. So there are choices being made to fit a desired condition. And then its projected from there.
F=ma does not have free parameters. Not that it matters since the existence of free parameters does preclude a model from being correct. Case in point…F=Gm1m2/r^2 holds irrespective of choices of F, m1, m2, or r and it has the free parameter G.
That is the argument I’m questioning.
Choices. That is a topic worthy of discussion. One thing non-programmers don’t realize is that choices are made for non-physical reasons. It’s not that they don’t know about or want to program the most physical model possible. It’s that they can’t. For example, we could theoretically solve convection and cloud formation explicitly without parameterization schemes. Models are capable of doing this. The problem is that it requires very high resolution models on the order of 1 meter. For example, a 50 km horizontal resolution with 50 vertical levels and a time step of 300 s (to comply with CFL) with convection and clouds parametrized is already 3 billion cell computations per day or about 100 quadrillion over the life of a 100 year simulation. If we wanted to deparameterize the convection and cloud physics it would be 0.001 km horizontal resolution with 10000 vertical levels and a time step of 6 ms (to comply with CFL). That is 7e25 cell computations per day or about 3e30 over the life of a 100 year simulation. So if takes 1 month to do the 100 year simulation with parameterization then it would take 3e12 years (or 200 times longer than the current age of the universe) without the parameterization. See the problem?
“F=ma does not have free parameters.”
Really? What is “a” in the case of a ball dropped from a height of 1′ above the surface of the earth?
Everyone accepted the reality of gravity (and pain) by ~ 3 years old.
…without realizing that models used to explain it have free parameters.
Experiments show that gravity exists. I am not aware of any experiments that show climate sensitivity exists, but it is a tunable parameter.
The existence of a climate sensitivity comes from the 1st law of thermodynamics ΔE = Ein – Eout and the heat capacity of a system ΔT = ΔE / m*c. BTW…for global circulation models climate sensitivity is usually not a free parameter since it can emerge from calculations.
Climate sensitivity is not a measured value. It is whatever it needs to be to make the climate model give the wanted answer. Therefore it is not a physical constant. It’s why it varies so much. It is free to be whatever it needs to be!
“It turns out that scientists cannot agree on what G is either.”
What it means is our ability to measure the value is limited. Different people get different results, just like two different people timing the period of a grandfather clock using a stopwatch calibrated in seconds.
That’s the entire basis of “measurement uncertainty”. The concept of which you don’t believe exists.
I was wrong.
You’re not better than that after all.
It means that no measurement is exact, uncertainty is the monarch of physical measurement.
The disagreements aren’t with the concept, they are with the margin where uncertainty is greatest. Look at your graph. Do you thing a torsion balance vs a pendulum will provide the same measurement? At what point do they disagree? You are trying to compare the lack of agreement with how to measure G, with claiming the concepts of relativity and gravitation are incorrect. An ill posed problem.
Get your head out mathematics and describe to the good folks here why these simple ideas have not been completely measured and determined with the technology we have today.
Here is a question to ponder. Why is a meter DEFINED by a rule using other physical concepts like speed and time, rather than an actual physical device to define the measurement length? It was at one time.
Statistical descriptions of observational data are effectively probability distributions. As such, there is not yet a deterministic theory of climate. Deterministic theories often employ constants, like E=mc^2.
You nailed it. /since there is a probability distribution it means there is an uncertainty interval associated with what ever value you pick for it. It’s what standard deviation is meant to describe. But, of course, you never EVER hear a climate scientists talk about the uncertainty in their projections.
one way to look at this problem is to determine if the boundary conditions of planetary atmosphere have any freedom at all. For example, it can be seen in any energy budget diagram that atmospheric absorbed radiation resembles 3/2 OLR. This suggests a geometric constraint. If the boundary conditions have no freedom then it is conceivable the system is deterministic. We are still floundering around in discovery at this time. No theory yet exists.
There are no free parameters in E=mc^2. The relationship between mass and energy is known exactly.
The *relationship* between mass and the gravitational constant is known as well. The only question is what the exact value of G actually is. It is the same with “c”. We don’t know it exactly either. REAL physical scientists know “c” has an uncertainty interval.
Your transparent attempt as dissembling is duly noted. “Relationship” doesn’t imply perfect knowledge of the associated factors in the relationship!
Yes, but the speed of light is now a “defined constant”, without an uncertainty internal:
https://physics.nist.gov/cgi-bin/cuu/Value?c|search_for=adopted_in!
You are right about G, it has an interval:
https://physics.nist.gov/cgi-bin/cuu/Value?bg|search_for=abbr_in!
ROFL!! Leave it to NIST to know the EXACT speed of light!
It is only defined for calculations. I don’t remember where I read that. You’ll notice is limited to the units digit. We simply can’t measure either time nor length to a sufficient resolution to determine an actual physical value, so a defined value is used so that everyone is using the same constant.
It’s the same with E=mc^2. We can’t measure E well enough to actually know how close either of the other two values actually are.
Pi is a constant as well, but its exact value can never be known. But that never stopped some people from talking in circles, did it?
ROFL!!
Pi is a fundamental constant. But it is not a free parameter. The reason is because its value can be determined numerically without experimentation. The fact that it is transcendental is irrelevant. BTW…the pun was great.
Just how do you think pi was discovered?
No, the point is that it can never be MEASURED without uncertainty. It’s like I said, we can calculate the mass of proton but has a proton in a non-moving state ever had it mass measured to verify the calculation.
Quit trying to mix up physical measurements with mathematical calculations. It is a losing argument because it is an ill-posed problem.
Exactly.
Maybe I’m not understanding your point. I’m talking about models with free parameters like global circulation models, the law of universal gravity, general relativity, the standard model, and the like. I’m not talking about the relationship between mass and energy which is known exactly.
I’m unaware of the so-called free parameters in physical theory. The physics and logical structure is described precisely using math, not statistical assumptions.
Free parameters are necessary variables in a model that have to be experimentally determined. There are many of them in various models spanning nearly all disciplines of science. G is but one example and appears in both Newton’s law of universal gravitation and Einstein’s field equations. G has to be experimentally determined therefore it is a free parameter whose value is tuned based observations. I mentioned it above because (I thought anyway) that no one would seriously challenge their physics-based nature.
no i think you are talking about calibration or line fitting. physical constants are not free.
I am not talking about calibration. That is a completely different concept.
bdgwx, you say:
You underestimate the complexity of an iterative model of a complex system compared to say determining G, the gravitational constant.
Suppose we have a simple formula that we know to be true, like say:
Gravitational Force = (Gravitational Constant × Mass of first object × Mass of the second object) / (Distance between the centre of two bodies)2.
The gravitational constant can be experimentally determined by knowing the two masses, the distance, and the force. And once we do that, we can check its veracity in a variety of ways.
Now consider a climate model. Here’s a bit of code regarding melt ponds
Without this bit of code, some of the melt ponds might never freeze, no matter how cold it got … gotta love that kind of physics, water that doesn’t freeze.
Now, the number -10 in that is a tunable parameter. We could set it to -9, or to -12.
But the problem is, that parameter only exists because the claimed “physics-based” model is going off the rails. All that parameter does is force the model back on the rails.
Now, we can tune that parameter to select the value that gives us the closest value to the real world, so I suppose in some sense it is “experimentally determined” … but that is profoundly and fundamentally different from determining G from an experiment.
Here’s another example. From Gavin Schmidt’s paper on the GISS Model E comes this quote:
In other words, the physics simulated in the ModelE won’t keep the modelworld in balance. So you simply turn the tuning knob and presto! It all works fine!
Now, is that the only way to nudge the model so that the TOA balance is correct? Of course not. It’s just the way they’ve chosen. And there’s no way to determine if it is correct.
In fact, that U00 tuning knob worked so well that they put in two more tuning knobs … plus another hard limit. From the code:
Note that the MAXCTOP is another tunable parameter. Without it, their “physics-based” model creates clouds with too high a cloud top pressure. Rather than figure out WHY it’s going off the rails, they just put a hard limit on the pressure …
Is MAXCTOP “experimentally determined”? Only in the sense that it makes the model work better … which again is NOT how G is “experimentally determined”. Because if the model actually worked, if it was a faithful and accurate representation of the physics involved, there would be no need for the MAXCTOP at all.
I certainly hope that you can see the difference between G and MAXCTOP …
w.
I’m not underestimating the complexity of anything. The point is that the existence of free parameters that must be experimentally determined (aka tuned) does not make a model unphysical. In fact, adding free parameters can make a model more physical. Don’t hear what I didn’t say. I didn’t say that free parameters necessarily make a model physical. They don’t. I didn’t say free parameters are desirable. They aren’t. I didn’t say free parameters are a panacea. They aren’t. I didn’t say that free parameters are immune from overfitting. They aren’t. I didn’t say that free parameters do not add complexity. They do. There is a billion things I didn’t say about free parameters that commenters are inevitably going to pin on me anyway so that they can cling to the misguided belief that models with free parameters are unphysical.
I quoted your damn words. You said:
“Free parameters are necessary variables in a model that have to be experimentally determined.”
Consider the free parameter in my example above that is used to make sure that water actually freezes.
So what we have is a model world where water doesn’t freeze … and in reponse, instead of fixing the model, they’ve put in a free parameter that forces it to freeze.
Perhaps you could explain just how that is a “necessary variable” … it’s only “necessary” because the model is not actually modeling the real world. It’s modeling a world where water doesn’t necessarily freeze no matter how cold it gets.
And that is totally and completely different from say G, where:
• The gravitational model describes the real world with extreme accuracy and
• We can use the model to EXPERIMENTALLY MEASURE the value of G to extreme accuracy, three decimal places or more.
Cavendish didn’t just guess at the value of G. He knew the other variable of mass1, mass2, the distance between them, and force exerted. This let him directly measure the constant G
On the other hand, there’s no way to use the GISS Model E to experimentally measure to three decimal places the value of the parameter that is needed to force water to freeze … it’s just a damn kludge used to cover up a giant hole in the model.
w.
Ok. I’ll do my best.
It looks reasonable to me.
You can see that the sea-ice thermodynamics module e-folds the melt ponds in 30 days when temperatures indicate melting conditions. But when freezing conditions are present the decay rate switches to an e-fold of 10 days. It is important to note that there is an accumulation step that increase mass within the pond but the slow decay is still necessary since the mass can decrease via discharge, contact, and brine fluctuations. Anyway, in freezing conditions it only takes 30 days for 95% of the melt pond to decay with no accumulation meaning it is a net 95% loss. The clamp clause does two things. 1) it resets the IEEE754 value to 0 so that lt and gt comparisons can be used (if desired) to highlight pixels with or without melt ponding correctly and 2) it hastens the pace of the freeze and constrains liquid mass to above -10 C only. The important point here is that the freeze is occurring without the clamp clause exactly as one would expect in freezing conditions.
And don’t hear what hasn’t been said. I did not say the above is a complete description of the model. It is not. I did not say that the model is perfect. It is not. I did not say the way ModelE does it is the best way. It almost certainly isn’t. I did not say that melt ponding cannot be handled explicitly via fundamental thermodynamic laws. It probably can. I didn’t say that it should be handled explicitly. That’s not always possible given the computational complexity. There is a lot of things I didn’t say that some commenters may want to pin on me in a misguided attempt to say the sea ice parameterization scheme in ModelE is unphysical.
I’m sorry, but that is not what is happening. The model is reducing melt pond mass in freezing conditions exactly as would be expected.
Yep.
Yep. And the keyword word here is “EXPERIMENTALLY” which makes G a free parameter.
Yea…he used an experiment to determine G. That’s the point. And now that we know G from his experiment (and others) we can use the model to make predictions of F, m1, m2, or r as long as we know 3 of the 4. And the more we experiment and refine the value of G the better our predictions become.
Nor do we need to. The model already freezes water without the need of the clamp clause and the -10 C value.
I wouldn’t call water freezing in freezing conditions a giant hole. And because I use clamp and guard clauses all of the time in scenarios similar to this I wouldn’t call it a kludge either. I don’t know. Maybe I have a different perspective because I’m a programmer.
“Yep. And the keyword word here is “EXPERIMENTALLY” which makes G a free parameter.”
You *REALLY* can’t tell the difference between a MEASURED value and an ASSUMED BUT UNMEASURED value?
The GUM goes into great detail about the use of “measurement models”. Each of the factors in the measurement model is MEASURED in order to determine an unmeasurable quantity. In the climate models many of the parameters are *not* measured, they are not even part of a measurement model where other factors are directly measured in order to determine the value of the needed parameter. They are *assumed* to be of a value needed to keep the model from blowing up or to make it give the pre-determined answer. *THAT* is a free parameter. Cloud coverage is not measured, a value for it is just picked out of thin air. THAT is a free parameter. A measured constant, whether measured directly or by using a measurement model, is *not* a free parameter.
And using better equipment to measure things is *NOT* “tuning” the value of the measurand. It is simply increasing the resolution of the measurement, it doesn’t affect the actual true value of the measurand in any way. There is no measured true value of a free parameter used as a fudge factor in a model.
Your “experimentation is tuning” is just plain sophistry at its finest.
“You can see that the sea-ice thermodynamics module e-folds the melt ponds in 30 days when temperatures indicate melting conditions. But when freezing conditions are present the decay rate switches to an e-fold of 10 days. It is important to note that there is an accumulation step that increase mass within the pond but the slow decay [etc]”
I’m not even sure what to say about that rant. I assume you’re towing the company line for people who you think will believe you and be able to dismiss the issue. It’ll probably end up at ScS as an official debunking.
But its quite sad really.
Edit: That’s not aimed at you Tim, obviously.
It was not meant as a rant. It was meant as a brief summary explanation of what ModelE is doing. And when I say brief I mean really brief. There is a lot of details involved.
I don’t ever expect people to believe an analysis because and only because it came from me. I’m wrong a lot. So it is always good to do your own analysis as well. I invite you to download the source code and review it yourself.
I’ve already read some of ModelE’s code. At the time I went looking for hard coding and to check out the cloud implementation.
You just don’t get it do you. To what decimal place does G have to be defined before you would call it a constant? How about the true speed of light? How many decimal places would need to be measured before you call it a constant? How about a given mass? How many decimal places are needed before it becomes a constant?
This is endemic in climate science. Clouds, who the hell knows! We’ll just make it a free parameter and make it what we need to arrive at an answer. Ocean currents, who the hell knows! We’ll just make it a free parameter and give it values that we need to get an answer!
Everything you bring up has been evaluated using the scientific method of proposing hypothesis and testing them experimentally. Clouds, oceans, convection have no hypothesis describing a functional relationship that can be tested experimentally. THAT MAKES THEM FREE PARAMETERS!
Actually, you are babbling. Have you ever read thru the document code of a climate model? I have. Do you understand the CFL numerical solution on PDF constraints. I did, and posted here years ago.
Not only have I read through the source code, but I have compiled it and run numerical models on my own machine. And yes, I’m well aware of the Courant-Friedricks-Lewy publication from 1928. No, it’s not a “numerical solution on PDF constraints”. It’s a necessary condition for a convergent solution of partial differential equations (PDEs). Specifically the CFL places constraints on spatial and temporal domains for discretized model meshes such that information does not propagate through a grid cell faster than the time step. If the CFL condition is violated then instability will develop in the numerical solver. I’m most familiar with WRF in which you specify the time step in namelist.input file. WRF recommends 6 s per km to comply with the CFL. Your mileage will vary depending on the specific model configuration or model type.
And what does any of that have to do with anything I’ve said above?
You *are* babbling. Physical constants are *not* “FREE” parameters. You can’t just pick a value for G in order to make a model come out the way you want it to. The same thing applies to the speed of light, “c”.
The value for those constants is determined through measurement and has an uncertainty interval. But it is *still* a fixed parameter, not a “free” one.
The tuning parameters in climate models *are* free parameters. You can choose whatever value you need in order to make the model output what you want.
You have not mentioned the word *uncertainty” in any of your posts trying to justify constants being “free” parameters. The concept of measurement uncertainty is so foreign to you that you can’t even speak the word.
my thinking is bdw is conflating statistical methods and the nature of things (physics).
Even at the very basis of climate 101 where OLR = S(1-albedo), the albedo is an ad-hoc free parameter with no physical basis. So too is the environmental lapse rate of 6.5K/km. They are chosen because it works. Climate science is packed full of ad-hoc free parameters which in my judgement are not free at all, nor are they constant. They all represent boundary conditions strictly constrained in steady-state. Strictly speaking, no physical theory exists until they can be computed a-priori from first principles. For some it is difficult to grasp the difference. Physics is about describing how nature actually works. Ad-hoc adjusting free parameters is about making things fit. In a physical theory there are no free tuning parameters, there are only fundamental constants.
Something isn’t clicking here. The concept of fundamental constants and free parameters are not mutually exclusive. Fundamental constants can be free parameters. Here are two examples that might help with understanding. The speed of light c is a fundamental constant that is not a free parameter when used in a model because its value is defined. c is not tuned to fit observations. However, the fine structure value α is a fundamental constant that is also a free parameter when used in a model because its value must be experimentally determined. α is tuned to fit observations.
That thinking is incorrect. Statistical methods can (and often are) used to tune free parameters. That does not mean that a statistical method is a type of free parameter or that a free parameter is a type of statistical method.
no please don’t do that. there are no free parameters in physical logic. There is no tuning or free parameters in nature. It is the axiom of physical science that there is a physical law underpinning all observables. tuning free parameters is the realm of probability space. It means we have not yet described the nature of the system, so we use crutches like tuning in attempts to make prediction in probability space.
I didn’t say there was.
Newton’s law of universal gravitation
Einstein’s field equations
Standard Model
Are those not descriptions of the nature of reality?
The factors in each of these is MEASURED, not tuned.
Take ECS for instance. It is neither calculated from first principles nor is it measured, either directly or indirectly. It is *assumed* to be whatever it needs to be in order to make the model in question give the desired result. It is a “free parameter”. That is why its value varies so greatly!
That is simply not the case for any of the constants in any of the examples you list. Values for the constants are not “assumed”, they are MEASURED.
the constants and terms in the logical structures of physical theory (equations) are not free.
Exactly.
G and α are obviously free parameters since their values can only be determined experimentally. Is Newtonian Mechanics and Standard Model not physical theory?
“ its value must be experimentally determined. α is tuned to fit observations.”
Do you not understand the cognitive dissonance you are exhibiting with this statement?
When you *MEASURE* a constant you are *NOT* tuning it. You are defining the constant within the measurement capabilities of your measuring system.
“a” is not tuned. “a” is defined through measurement. A free parameter isn’t measured, it is ASSUMED. It is assumed to be whatever you need it to be – just like in the climate models.
You pretty much nailed it with: “Physics is about describing how nature actually works. Ad-hoc adjusting free parameters is about making things fit.”
People like bdgwx just can’t get it into their minds that a constant can have an uncertainty interval because human measurement systems are not perfect with infinite resolution. So when people get different values it seems to someone that doesn’t understand uncertainty that it isn’t a fixed constant but a tuning parameter where you can pick different values willy-nilly to, as you put it, make things fit.
G having been found to possibly vary in space or time is suggestive there is a problem with Newtonian physics. Whereupon there must be a deeper formulation of G. This technicality, however, is irrelevant to the far more elementary definitions here. Generally statisticians do not cross over into physical theory, and vice versa. Statisticians work with statistical models and theory, Physicists work with physical models and theory. It is two separate disciplines that share some commonalities in notation.
The speed of light is the speed of light, it is not a probability function. It may have an uncertainty interval because human measuring systems are not perfect but it still has one true value. It is a human failing that we do not know that one true value.
yes it is a fundamental aspect of our universe. The discovery of physical constants allows for many fascinating insights.
We already know there are problems with Newtonian physics and it has nothing to do with G.
Maybe…maybe not. But that’s irrelevant in this discussion. The question is…does the existence of a free parameter in a model like Newton’s law of universal gravitation, Einstein’s field equations, the Standard Model, etc. necessarily mean those models are unphysical? Willis says yes. I say no.
this is getting silly and circular. Stick to your curve fitting projections.
Unfortunately we all have to stick to fitting models to observations since many of them have free parameters. It sucks. But science has not yet progressed beyond them yet. There is a Nobel Prize up for grabs to the first person who can show that G, α, or many of the other fundamental constants emerges from a calculation or fundamental principal as opposed to being estimated by experiment. Until then…continuing to tune free parameters is our future.
yes the natural world is full of fixed parameters. without them physics could not exist as we know it.
How do you calculate the speed of light? Does it fall out of some calculation somewhere? Or is it’s value MEASURED and then used to calculate other parameters? What calculation could possibly define a constant? Constants are *measured*, not calculated. Pi was discovered by measuring the circumference of circles and their diameter and comparing them. It didn’t just fall out of some calculation of a different parameter!
Your sophistry is tiring and boring. It just shows that you know you are wrong in your claim. You are just being stubborn and bullheaded in trying to define parameters that can take on any value as being the same thing as a physical constant – and vice versa,
Oh, why would you besmear anyone’s professional reputation by fitting them up with a much- devalued Nobel Prize as selected by the current woke judging panels?
bdgwx writes “Unfortunately we all have to stick to fitting models to observations since many of them have free parameters. It sucks.”
It makes them not based in physics. And the fact they project a climate state and then step again from that as an initial state means they have no predictive value at all.
That’s the debate. Are many of our scientific models like Newton’s law of universal gravitation, Einstein’s field equations, the standard model, etc. not based on physics because they have free parameters that must be determined experimentally? I say yes.
You dont understand the nature of the fit in the models. Its not a matter of “fitting” clouds to get better clouds…its fitting clouds to get a better unrelated quantity, TOA radiative unbalance.
Personally I think you’re off on a tangent with your argument. Fitted quantities are well established to not have predictive capability. G isn’t fitted, its measured. It applies consistently across time and space.
Calling the refinement of a measurement “fitting”, is your own problem.
bdgwx simply can’t admit, not even to himself, that measurement is not the same thing as assuming. Clouds in the climate models take on an “assumed” value, not a measured value. You can assume whatever you need to get the output you want. Physical constants are not assumed at all, they are *MEASURED*. No tuning, no assumptions to make a model work out the “right” conclusion. For him to admit to the difference would require him to also abandon the climate models as unfit for purpose. That’s never going to happen – thus the cognitive dissonance.
In the case of G, the relationship between mass and distance is postulated as a theory, experimented with, and tested and the constant G derived from measurements.
G is solid even if refinement of its value continues over time with better measurements.
However if G (and its related formula) was then further included in a fitted equation to correlate with say earthquakes based on nearby mountains and the tides to describe their frequency or likelihood then that would be an unphysical use with no predictive power.
This is the equivalent of what the GCMs are doing by relating clouds to the TOA.
I’ll paste this again from Wikipedia.
“general relativity reduces to Newtonian gravity”
That doesn’t mean G is not a proper constant under the correct conditions of small potentials (tall buildings) and low velocities (<<< c).
‘bdgwx’ in totally unable to understand the process of making a hypothesis, developing a functional relationship, and designing experiments to test the relationship. As Willis has tried to point out, G has been measured, with experiments. GCM’s do have free parameters that are used to counteract the inability to have functional relationships that accurately describe what occurs in the atmosphere. Those parameters were not built from functional relationships that can be physically tested to determine their physical value. They are used to keep the model “within the rails” as Willis described it.
hi Jim,
yes tuning is an empirical method which can lead to development of rational theory. However, in many types of applications there is no need. For example, hydrologists use little empirical tricks to correlate precipitation intensity to catchment runoff. The runoff volume is tuned to the catchment of interest based on historical streamflows there. It works in many applied sciences. Weather forecasting and climate are using probability distributions to represent the parameter space. It kinda sorta works to force desired results or outcomes that appear to relate to reality. However, this is not a replacement for a testable physical theory which give new insights into nature. An empirical relation is not a physical theory. It should be noted that GCMs are the height of empiricism and may offer clues into the nature of things. However, they are likely the worst possible tool to depict a physical theory of planetary atmosphere. It is therefore unsurprising no theory exists. Very few even seem to recognize this, despite many shouting that the science is somehow settled. Often, the best insights into nature are depicted with a foundational set of straightforward physical relations. One can then attempt a GCM type exercise. For now in the rush for political persuasion the GCMs are floundering without strict physical constraint, because no foundational theory exists. Instead, they rely on ad-hoc parametrization.
It should be noted additionally that many strict radiative constraints are existing already in the literature. As discussed in Goody 1964, Chamberlain 1978, Houghton 1997, Hartmann 1994, Mihalas 1978, Stephens and Greenwald 1991, and so on. But they are somehow ignored, still.
It is plain to see that radiative boundary conditions have strict constraints, and are not free at all. In my view the only remaining freedom for theoretical investigation is the ratio of total incoming solar to the partitioning of surface turbulent flux. Everything else, from albedo to surface net radiation, emerges directly from theory. It is all largely a function of surface sensible heat H.
Surface total radiation = 2OLR + H. So, the primary question is the physical control on H, naturally.
That is great. I would love to see a model, especially one devoid of free parameters, that produces better results than a GCM. Would you mind posting a link to one that provides monthly global average temperature predictions from 1880 to 2020 and which has a root mean squared difference from observations that are significantly less than 0.1 C?
The most profound and fundamental insights are depicted with simple models, in thermodynamics these are referred to as zero dimensional energy balance models.
If one becomes stuck in GCM land, blinded by their apparent sophistication, it is easy to lose perspective. It is an audacious endeavor, and worthy of effort.
However, it is an inappropriate tool for studying the foundational constraints. The goal of the physicist isn’t to fit data or to tune to minimum RMSE. The goal is to gain insights into how nature works.
People tend to get lost myopically chasing their tails in GCM model land, with endless tuning.
Yeah. You made that clear above. So what model other than a GCM is a better tool for depicting the planetary atmosphere? I’d like to be able to download the monthly global average temperature data from 1880 to 2020 so that I can compare it with observations to see just how much better it is.
Interesting. So scientists should gain insights into how nature works by using a model that doesn’t necessarily need to fit the observations or minimize the RMSE? I certainly don’t want to misrepresent you here. That’s why I’m asking.
It is the statisticians goal to fit data and minimize residuals.
It is the physicists goal to describe the logical structures of physical systems.
For example, a physical description is normally informed by observation, such as the approximate numerical depictions in a Trenberth diagram inform the following relation.
Surface total radiation = 2OLR + H
This gives insight into the logical structure of atmosphere.
Pure gaseous radiative constraint can be depicted using clear sky CERES data, such that:
Surface net radiation = OLR/2
This informs physical understanding. It is food for thought, if you will.
“The goal of the physicist isn’t to fit data or to tune to minimum RMSE.”
While happily ignoring the actual measurement uncertainty in your data.
Let’s be accurate. From Wikipedia:
In situations where either dimensionless parameter is large, then general relativity must be used to describe the system. General relativity reduces to Newtonian gravity in the limit of small potential and low velocities, so Newton’s law of gravitation is often said to be the low-gravity limit of general relativity.
Does that mean G is not a constant in “in the limit of small potential and low velocities”? No, it does not. It also means that G is not a “free parameter” that can be changed at will to make further calculations. Newtons Law of Gravitation is a functional relationship under the correct conditions.
The hypothesis of relativity was proven by experiment. The functional relationships that Einstein developed were proven to be correct. You apparently are still having a hard time reconciling the difference between functional relationships and physical measurements even after all the discussions on this subject.
Yet E has an uncertainty associated with it. Although we can calculate the mass of a proton, can we physically measure it? Relativity comes into play in determining mass. Has anyone ever measured the mass of a non-moving proton?
To how many decimal points can we measure a mass? Do you think Avogadro’s number is exact? Do we know the speed of light to an infinite number of decimal places? Can we measure the length of time with a sufficient number of decimal places to determine the speed of light?
Golly gee, as Private Gomer Pyle would say, UNCERTAINTY seems to be the rule!
The G constant is only there to match reality to out measurement system kilogrammes etc. We could come up with a system with no constant, because the system exactly matched nature, but that might be a bit inconvenient. It would be very interesting to see if those units matched every natural thing with constants, but it would be a lot of work to develop it!
That’s the ultimate goal…a theory of everything with no free parameters. It’s a noble goal for sure. Unfortunately it has eluded scientists to the point that there are numerous free parameters entrenched in science now. The Standard Model alone has 19 free parameters that must be tuned via experimentation.
You do not “tune” physical parameters, you MEASURE them, You refine your measurements as much as you can with the equipment available to you but they still carry an uncertainty interval with them. Refining your measurement is not “tuning” it, it is measuring with higher accuracy and resolution. It is *not* a free parameter.
Face it, you know you are wrong and now you are trying to come up with a definition for “free parameter” that you can use as an excuse to claim you were correct that physical constants are free parameters.
Refining measurements is not tuning. Tuning is not refining measurements. You are trying now to make them the same thing.
bdgwx, Newton was perfectly right, according to the science of his day. At least, unlike Einstein, his work still has value for the common working man. However, there is a thing called “classicism”. It requires you to base all new knowledge upon known and RESPECTED predecessors, come hell or high water.
Any deviation from adherence to classicism automatically classifies all you have to say as ‘pseudoscience’.
Now you know why the plasmatic universe theory is a joke, while virology and vaccines are worshipped as actual science.
P.S. An electric universe would get rid of that G-parameter…
“I thought that the answer to that was that all the models are tuned to the give good agreement to historical temperatures.”
Climate Alarmists wouldn’t know “historical temperatures” if they walked up and bit them on the butt.
It is not a historical temperature record if the record does not show the Early Twentieth Century as being as warm as today. These bogus, bastardized Hockey Stick charts used here do not represent reality. What is correlation with non-reality? That’s what we have here.
You mean the Divination of ancient tree rings isn’t accurate to several decimal places?
Or at all.
I tried it with an old tree we had to cut down in our back garden. The answer was 47 🙂
That made me laugh! 🙂
So strange to look at maps where the Pacific ocean is centered and Europe is split in half. On any other news site it would have tripped a “something’s fishy” alarm.
Willis, it appears you caught them red-handed, yet again.
I won’t hold my breath waiting for the modellers chastened words of apology for their obviously unintentional errors in predicting Armageddon when in the real world everything just gets better and better.
It is tempting to smugly “tell California” that other states will happily accept businesses fleeing high cost, but the businesses come with people. The people bring their politics and expectations with them. We are all California?
California politicians loved to repeat the phrase: “California leads the nation”. Recently though, they’ve stopped as the exodus continues to lead the nation as well.
I’ve heard that those who move to Texas get Texified pretty quick!
(E2) https://youtu.be/3MOy6Z_UP7c
(E4) https://youtu.be/WxQVUNPeWnc
(E5) https://youtu.be/xsImFkLdE4A
When they get to the bottom, they’ll stop “leading”. Won’t be too long now.
Great work, WE. I have a pretty strong hunch about the underlying causes of figures 7 and 8.
We know the models are tuned to best hindcast 30 years. That hindcast is a CMIP written requirement as submission 1. When I went thru the guts of a CMIP3 model documentation, and then the tuning papers described in a previous guest post here (The trouble with climate models) it was apparent that two of the most common hindcast tuning variables are aerosols and cloud fraction. Both reflect incoming solar so increases reduce surface absorbed solar to ‘cool’ the models to ‘match’ reality as reflected by BEST. Explains fig. 7.
But the tune stops at the beginning of the forecast run. Without the tune they would have run hot in hindcast, so do automatically in forecast. Fig 8 explains one reason why—steadily increasing surface absorbed solar.
The models are doing exactly what they were intended to do. Scare people into thinking they model ‘reality’ and that ‘reality’ is ever increasing global warming. Even though Fig 8 forecast probably isn’t close to solar surface reality. You would think modelers themselves would check this sort of thing out, but that is apparently not their job.
Whatever satisfies the agenda is the “correct” answer.
(That’s observed practice, not speculation)
Agree. Several irrefutable model based examples.
These are not coincidences, but rather narrative driven scientific sloth.
“All CMIP6 except INM CM5 produce a tropical troposphere hotspot that does not exist in reality.”
I noticed on a chart the other day that the Russian model was closest to the actual temperature readings, but was still a little bit high compared to observations.
The Russian model uses a far lower figure for climate sensitivity but still is subject to the same limitations as other climate models.
“Both reflect incoming solar so increases reduce surface absorbed solar to ‘cool’ the models to ‘match’ reality as reflected by BEST”
“Reality as reflected by BEST” is actually a False Reality.
And that’s how the climate alarmists match the bogus models to the bogus, bastardized “temperature” record.
They start out with a bastardized temperature record and then tune their models to duplicate it.
Then, the climate alarmists count on people not noticing that the actual wriiten, historical temperature records do not have a “hotter and hotter and hotter” temperature profile the way the bastardized Hockey Sticks chart do, instead they have a benign temperature profile where we are not experiencing unprecedented warming today because it was just as warm as today in the recent past.
There’s a whole lot of pretending going on here in skeptic circles. Pretending Bogus Hockey Sticks represent reality. The written record is clear. And it’s not “hotter and hotter”.
It is kinda like trying to argue a legal case in a court with a panel of corrupt judges. You can have a fantastic case, evidence, tons of sworn testimony, meticulous search of all related case law. But it doesn’t matter.
While there is great risk in biased AI, truthful AI could solve a lot of dishonest agenda modeling and associated bias in reporting. Or maybe Willis can just look into it and get decent coverage on the same. Thanks also to Anthony for saving us during the model crusades.
“While there is great risk in biased AI, truthful AI could solve a lot of dishonest agenda modeling and associated bias in reporting.”
The Trick will be how to tell which AI is honest and which one is not. So humans are going to end up having to do their own research anyway, by fact-checking everything the AI says. The AI has already been caught telling lies, even making up lies. So good luck with AI. Don’t bet the farm on what it says.
Willis: as always, a very interesting exposé. Under the first Figure 7 (there are two) you say “The models have done a decent job of emulating the major changes in the historical temperature. (It does bring up the question of how different models with widely differing climate sensitivities can all hindcast the temperature so well…”
I’ve found it’s always easier to accurately predict the past than the future. Cue Yogi Berra.
“Historical temperatuers” is being used incorrectly in this article.
What we really have here is computer-generated “temperatures”, that do not reflect the written, historical temperature records of the world.
Willis,
Thank you for again posting a most interesting and intellectually stimulating analysis, here one which examines surface-absorbed solar radiation versus surface temperatures both for land and for oceans.
One thing that is, to me, very curious are your plots given in Figures 4 and Figure 6 showing “hot” spots that exist in the Pacific and Atlantic oceans in the tropical zone south of the equator and which do not have counterparts in the tropical region in either ocean north of the equator.
The stated time periods of the data portrayed, about 22 years, would be sufficient to average in at least a couple of PDO cycles and at least one AMO cycle (but I thought AMOs were confined to the Atlantic north of the equator), even if these were somehow relevant to surface absorbed solar radiation.
Could it be that an abundance of near-surface plankton just south of the equator is the cause of these “anomalies” in both oceans? Could these south-vs-north differences be related to geographical differences in percentage of average cloud coverage over the stated intervals?
Your thoughts on this are most welcome.
Build new fossil fuel and nuclear generators and remove all wind and solar from the grid.
Willis,
So .21 degrees per watt SW. Wouldn’t this cause one to think that the 3 watts calculated by Happer and others for a doubling of CO2 would result in 0.63 degrees of warming ?…..a fraction of IPCC predictions….or am I missing something obvious ?….only functioning on 1 coffee so far today.
Unknown. Too many other variables involved.
w.
The change in the sun’s luminosity has no apparent effect on the the earth’s energy imbalance. This would suggest that the climate sensitivity is quite close to zero. In fact, it appears to be indistinguishable from random noise.
Assuming the graph of solar radiation absorption is accurate, we see a sort of negative correlation between absorption at the surface and “average global temperatures”. This implies two things: the change in solar absorption at the surface and the modest increase in temperatures would have to be an amazing coincidence if they aren’t directly related, and if they are, the late 20th century increase in industrial particulates entering the atmosphere is probably screening out some sunlight, but is substantially reducing radiative cooling at night, especially during colder months as has been observed.
-Johne
Separate observation concerning figures 3 and 6. 3 shows no warming with increased insolation above 210. 6 shows where, the tropics. I believe the combination is yet another elegant way to demonstrate your tropical Tstorm thermoregulation hypothesis to be correct.
Even tho Fort Lauderdale is not technically the tropics, it is most definitely subtropical. We seldom have Tstorms in the winter (thermoregulation not needed), yet we have them almost every day in summer when we reach the fig 3 temperature threshold—as is happening here right now. It was 86F outside an hour ago when I walked the dog (we could see the Tstorm coming from the Everglades). Now it is 77F. Air temperatures, not sea surface or here Everglades water surface.
True indeed,Rud.
Thanks,
w.
Willis’ tropical Tstorm thermoregulation hypothesis seems irrefutable, and that causes me to marvel at the following disasterous forecasts for the following tropical coastal towns in Australia.
Perhaps they sometimes wonder why there have been no changes to maximum temperatures so far:
BOM data charts from 1941 by month, in 30 year segments. (The historical data plot is overlayed by the current data plot, so it takes a bit of focus to see it peeping out here and there!)link
Could it be there is a seasonal artefact in the data? Earth is warmer during NH summer and colder in NH winter. Equally there might less clouds in NH summer, as opposed to winter. What would happen if you give the NH summer months a different color in this chart?
There has to be. Summer in either hemisphere is when that hemisphere tilts slightly closer to the sun, giving slightly more energy concentration per m2, winter is when it is tilted slightly further away, giving less energy per m2. Given the averages and outliers on the charts Willis has shown, there must be a ‘residual energy’ factor – a certain amount of retained heat that only changes slowly over the seasons but would affect the change in temperatures from more solar energy. For example, I would imagine it would take less energy to raise temperature 1C within the ‘average’ range of temperatures than it would to raise the temperature 1C as an outlier – that is, from 12C to 13C might take less energy than to raise it from 28C to 29C?
Good question, E. I just checked looking at just the NH data. It gives the same answer.
w.
Ah that old chestnut..
Vuk
Is the sun source of periodic variability of global temperatures?
ChatGPT
Yes, the Sun is one of the primary sources of periodic variability in global temperatures. Solar radiation, specifically electromagnetic radiation emitted by the Sun, plays a crucial role in Earth’s climate system. The Sun’s energy output is not constant but varies over time due to natural processes occurring within the Sun.
One of the well-known solar variations is the 11-year solar cycle, also known as the sunspot cycle. Sunspots are dark regions on the Sun’s surface that correspond to areas of intense magnetic activity. The number of sunspots varies over time, reaching a maximum (solar maximum) and minimum (solar minimum) roughly every 11 years. During solar maximum, the Sun emits more energy, while during solar minimum, the energy output decreases. This variation in solar output can influence Earth’s climate and temperature patterns.
Although the exact mechanisms linking solar variability to climate change are still being studied, scientists have found correlations between solar activity and certain climate indicators. For example, during periods of high solar activity, there tends to be a slight increase in the amount of solar radiation reaching the Earth, which can have a small warming effect.
However, the overall influence of solar variability on global temperature is relatively small compared to other factors such as greenhouse gas concentrations.
My (sarcastic) answer would have been much less informative.
(Thank goodness I do hardware instead)
Vuk
Is the sun source of periodic variability of global temperatures?
KevinM
Yea, we call them seasons.
The great irony of AI might be its use of English language user interfaces when the programmers probably speak some other languages first. I remember New England USA engineering TAs used to be grad students, and not many of the grad students had been born in New England USA.
One consequence of programming language building itself around spoken language and interaction between the two might be a future where “regular people” start to speak c-like English in 20 years?
Ironically (there again) … the easiest way to verify might be by asking AI in 20 years.
We also call them ‘days’ – from a certain perspective the changes form a fractal pattern. The change in seasons is a fractal of the daily change, warm period to cold a fractal of the seasons and glacial/interglacial a fractal pattern of that change. I find it hilarious that climate enthusiasts acknowledge a cyclical change in days, seasons and glacials yet scream bloody murder when you dare to suggest a cyclical pattern in multidecadal warm/cold periods!
Richard, the days, seasons, and glacials are regular and are tied to understood solar system variables.
Multidecadal warm/cold periods, on the other hand, not so much … the AMO is perhaps the most regular, but we only have data for about two cycles, and the early data is poor.
And the length of it is ~33 years … why 33 and not 42, or 25?
Best to you and yours,
w.
“However, the overall influence of solar variability on global temperature is relatively small compared to other factors such as greenhouse gas concentrations.”
ChatGPT shows its bias. ChatGPT is spewing the Party line. A more perfect propaganda effort: It must be true, ChatGPt said so.
Presumably they “adjusted” the absorbed solar radiation as a way of matching actual historical temperatures. What happens to modeled future temperatures if you continue reducing solar radiation at the same rate? We may just keep it below the dreaded 1.5˚C rise about the time absorbed solar radiation hits zero.
Or I wonder what might happen if the rising temperatures were only an artefact of the UHI and, outside the towns and cities, temperatures were actually falling, more or less, in line with solar radiation! Heresy!
A couple of comments:
The surface is the warmest spot of the climate system. Heat flows from there. Hence the surface temperature is important. The notion that oceans can be heated top down is unsupported by fact. Heating the surface simply accelerates the water cycle and actually cools the deep oceans as the high latitude cool, less saline water flows toward the rising equatorial water:
.
Northern oceans are retaining more heat because the summer water cycle in the NH is slowing down as the NH land warm ups in response to increasing peak solar intensity. In the current warming cycle, the June solar intensity at 40N is up by 2W/m^2 with 23W/m^2 yet to come. There will be a lot more warming before snowfall overtakes snow melt again.
The ocean surface temperature has very little change night or day. Not much more than 1C:
stack5s156e_20100610_20100618_hr__tt_eps281t_2023061918.png
Ten days in June when the Pacific moored buoys were still being maintained.
“Deserts are cold at night no matter how hot they get during the day.”
I don’t know why this is always taken as an uncontested fact. Hot deserts aren’t usually “cold” at night after hot days. Here are climatological averages for Riyadh:
MonthHigh / Low(°F) Rain
June108° / 82° 0 days
July110° / 84° 0 days
August110° / 84° 0 days
September105° / 78° 0 days
I remember going TDY to Vandenberg AFB one summer. Traveling between Colorado and California, I stayed overnight in billeting on Nellis AFB outside of Las Vegas. It was 100⁰ at midnight, and keeping the A/C on full blast all night only kept my room temperature around 80⁰. It was miserable, and needless to say, I didn’t sleep well that night.
High deserts far above sea level with low humidity do in fact have large diurnal temperature variations, but those are usually considered “cold” deserts (Bwk and Bsk).
Figure 4 would be more interesting to split the response between water and land. Most of the heat going into water produces water vapour so there is lower response in temperature to absorbed solar.
California doesn’t fool around. But then neither does Colorado. This just in from Colorado Public Radio’s web site:
Shocked.
One day there will be a federal, state, city and township – and parking meter – holiday for the final debunking of the global warming myth. We will be free.
“And how come I’m the guy who notices these things and not the good folks running the models or the people at CMIP5?”
Perhaps they are using Mike’s Nature trick to hide the decline
Zhang and Rossow’s paper “Global Radiative Flux Profile Data Set: Revised and Extended” showed that downwelling shortwave radiation has been increasing since their measurements begun, early 1980’s.
According to this paper, downwelling shortwave radiation has increased by about 3 watts per squ m.
Using Willis’s calculations this means a warming of .6C since early 1980’s. There is also a cooling from an increase in outgoing longwave radiation.
UAH global temperature data set shows a warming since the early 1980’s of about .5C.
Thats pretty close.
Brilliant as always!
Explains temperature development in Sweden due to increase in sun due to less clouds and the lack of summer response (over 220 W/m2) Your figure 3.
Is this just a coincidence? https://www.bbc.co.uk/news/science-environment-65948544
Or this
both taken in the uk
the sunspots three days ago
the prominence three weeks ago.
Hi Willis. Can you tell me where you got the temperature data for Figures 1, 2 & 3, please?
I’ve used a Stefan-Boltzmann conversion of the CERES upwelling LW data. I’ve checked it against the other surface datasets, and it agrees with them about as well as they agree with each other. So I use it by preference because it is in agreement with the other CERES datasets.
For example, here’s CERES cloud radiative effect scatterplotted against first CERES and then Berkeley Earth surface temperatures.
For the kinds of anayses that I’m doing, the small differences are immaterial.
Regards.
w.
Points arising:
1/ The Sputnik says absorbed solar is decreasing.
How it knows that is, short of StarTrekTechnology = a mystery. Unless it’s doing a really simple Albedo measurement.
So if we go with that, NASA are saying that Earth is getting ‘lighter in colour‘
Well that’s easy to do:
So which is it?
You know which one I go for as it explains rising sealevels, reducing length of NH growing season hence = rising CO₂
i.e Soil Erosion.
We all know that ‘bone dry’ fields. gardens, roads, yards, roofs and bone-dry everythings are lighter in colour than ‘wet things’
That explains the rising temps dry things don’t have the ‘heat mass/inertia’ that water gives things – so when the sun comes out the heat faster and so to higher temps.
It’s well known about – the really clever people call it “Aridification”
I call it Desertification.
2/ To my mind it’s more obvious than a ReallyObviousThing that NASA are trying to implicate green house gases as the warming agent.
C’mon, how can absorbed solar be going down and temps go up. The thing (Earth) that’s warming MUST be getting energy from somewhere and GreenHouseGases are the only possible answer in these modern times.
3/ Willis does us a nice little graph (his Fig. 3)
How that divvies up into its 3 sections is quite simple.
The left hand part, the really steep bit at less that 50W is where water has become ice.
So we’re right back onto the Albedo trail – snow/ice have an Albedo of 80% – compared dark soil at 10% and deep clear water at 6% (normal incidence)
The middle bit made a nice comparison with my ‘all water world’ Excel model.
In my model and for latitudes below where water freezes (60° latitude) – the average temperature response in my model is 0.18 degrees Celsius per Watt
And the sensitivity rises above 50 degs latitude but not as much as my simple model doesn’t allow water to freeze.
The right hand part where things level off is seen in my model and in The Real World and as noted many times by TheManHimself = that deep water never gets above 31°Celsius.
This drops straight out of a Stefan radiation calculation.
i.e. That on a spherical object at the equator, the average solar power over a 12 hour day is (after Albedo of 0.06) = 914Watts/m²
To be in equilibrium, the water must radiate away half that much over 24 hours, via an emissivity of (say) 0.95
(you work it out)
4/ The really awful bit coming out of the 0.2°C per Watt of absorbed energy.
Everyone is gonna leap up and down, especially his Ludship Monckton and yelling about Equilibrium Climate Sensitivity (ECS) – and There It Is – Willis and The Sputnik have found it.
The trouble with ECS is a guy called Carnot and = why Climate Science is such an awful mess.
Because all Watts per Square Metre are not equal.
It is that when Energy heats an object/substance/gas, it is imparting mechanical energy to the molecules/atoms of that substance = the very definition of any/all Heat Engines
Hence, to calculate a heating effect, you need to account for the temperature of where the energy is coming from and the temperature of the object absorbing that energy.
For the sun, the calculation is almost trivial because the sun is so hot = the Carnot Efficiency is over 95% – meaning that for objects on Earth almost all the solar energy becomes = heat (higher temperature)
BUT EVEN THEN – where is that 95% figure in any of climate science? It’s important when we’re quibbling about a few fractions of a Watt
But for the supposed Green House Gases – they are not coming from an object of 5,700 Kelvin, far from it.
Run an example, as per Tyndall’s experiment.
Tyndall used boiling water as his energy source and his temp reference and why not, it makes his experiment repeatable
Tyndall’s absorbing object was the gas in his test chamber and it would have been at what – maybe 20°Celsius?
So the Carnot Efficiency then would have been 21% in Tyndall’s experiment using boiling water
So to get the same heating effect from as we do from El Sol, we would need 95/21 = 4.5 times more Watts per m²
IOW To get the same effect using boiling water rather than El Sol would require 4.5×0.21 watts per m²
call that = 1 Watt/m²
And that’s using boiling water – NOTHING in the atmosphere is that hot
Lets run a Santa Ana wind over a glacier
Santa Ana = 35°C
Glacier = minus 10°C
Carnot efficiency would be then = 14%
i.e. The wind really got it’s work cut out to inflict anything much upon the glacier
Foe Climate Science it gets impossibly worse.
Because of Lapse Rate meaning all parts of the atmosphere are always colder than anywhere below that part, the Carnot efficiency becomes negative
Lapse Rate and Carnot make an utter mockery of the GreenHouseGasEffect
edit/reply to PS
Back to my Excel Water World
I divided the globe in strips of equal latitude then worked out some sines and cosines to get the average (over 12 hour) absorbed energy, hence what needs to be radiated away over 24 hours.
You can do this because ‘water stores energy’ – the temp drop of 100 cubic metre of water overnight would be about 0.05Celsius
I then did a brute force integration, working out TempxArea, summing those and dividing by the total area
I got a global average temp of 10.5Celsius – using an albedo (water) of 0.06 and emissivity of 0.95
(My earth has no atmosphere = no clouds and why not, CliSci says earth;s atmosphere is transparent and you don’t get more transparent than = not there)
The really laughable thing is when I ask Excel for a simple average of the Watts/m² power – it tells me 303 Watts/m²
Doing it properly gets a warmer Earth (10.5C vs minus18C) for less power (303 vs 340)
What a total joke it all is.
Reply reply…
Hopefully it dawns on why clouds(esp), supposedly ‘reflecting heat’ back to the ground/earth/surface would have no effect.
And why putting a mirror or shiny foil near warm objects does not stop them cooling.
It comes straight out of Carnot that the warm object, via the mirror/foil/clouds is radiating upon itself – so the source temperature and the absorbing temperature are the same.
The Carnot Efficiency thus becomes= Zero
i.e. The object radiates energy and that energy can not come back
Josef Stefan said exactly that when he that that “Objects radiate according to their own selves – what any other object anywhere does is irrelevant”
Sorry, all that shiny stuff inside your Thermos Flask DOES NOT work by ‘reflecting heat’
Clouds in the sky don’t do it neither.
They can and do reflect solar energy because : It is hot so it can be reflected onto a cooler object and be absorbed there.
Clouds on Earth do not affect the temperature of the sun – the GHGE says they should
The optimist thinks that he is living in the best of all possible worlds. The pessimist fears that he may well be right.
Low temperatures in the upper stratosphere over the southern polar circle.
Willis, if I am understanding your writings correctly, adsorbed solar is inversely proportional to cloudiness as in Fig 1 and cloudiness has been declining in recent decades, hence higher temperature. If cloudiness is declining, then so must be humidity and climat4you.com shows declining humidity at all tropospheric elevations in recent decades. So far, so good. But why is humidity declining? Isn’t CO2 supposed to increase temperature and humidity?
The Earth’s troposphere is very thin. Water vapor decreases and causes strong surface heating, especially when the jet stream weakens and stationary highs form.
The opposite is true in winter, the surface temperature in stationary highs decreases rapidly.
I think Willis that you have made an absoloutely huge discovery, for which all credit is due. Why does the surface appear to absorb less energy at higher temperatures? Your plot (long-term averages of both temperature and absorbed solar radiation) is the key, because it shows what happens when there is more water vapour in the atmosphere, water is the temperature controller for the Earth. It would be interesting to do a similar plot of average humidity against energy absorption, or particularly a surface 3d plot, and I think you have the evidence against all of the modellers. They forgot the water!
Clouds control the temperature of the Earth. The amount of clouds depends on the amount of sunlight striking the wetted surfaces of the planet. Why the control hasn’t been tighter than a degree in the last century is what is elusive.
Why does my old car wander its way down the road? Why the control isn’t tighter is elusive. No control system is perfect, there is always some slop due to tolerances and hysteresis. The earth’s temperature is much the same.
For those confusilated by bdgwx’s bizarre idea that everything is a “free parameter”, I can do no better than to cite Freeman Dyson on the subject. Click on the image to enlarge it.
w.
I think you have me confused with someone else. That is not my idea nor do I support it.
“I think you have me confused with someone else. That is not my idea nor do I support it.”
Really? Then why did you say:
” If it can only be determined via experimentation then it is a free parameter.”
*Everything* is determined by experimentation. Measuring the length of a 2″x4″ board is an experiment. There are many ways to actually do it. You can use a measuring tape. You can use geometry/trigonometry by comparing it to the height of a known (ask a Boy Scout how to measure the height of a tree). You can stand it next to something whose height is known. These all result in experimental results.
But the height of the board is *not* just ASSUMED hoping it will fit in the stud wall you are building!
Willis says
“ surprisingly, in areas where average absorbed solar is above about 210 W/m2, increasing the absorbed sunlight doesn’t warm the surface much at all. The average response in the colored areas shown below is 0.03°C per W/m2. Go figure. Here are those locations.”
My impression is that those areas are ocean, and so, depending on the spectral W/L of the surface TSI, some of said energy goes below the surface, and is transported to other regions before returning to the surface. A change is solar spectrum here could have very slow but long term effects in earth’s energy budget.
I have speculated that this help’s account for why the oceans are, on average, warmer then the land surface.
In the likely futile quest to get bgdwx to wake up, let’s consider this note about a tunable parameter in the GISS ModelE climate model:
So when temperatures are not “reasonable”, which is to say they exceed some tunable “boundary temperatures” parameter, they are simply reset back to the temperature of the previous timestep.
How do we know that these “boundary temperatures” are tunable parameters?
In part, because they have no counterpart in nature.
They don’t exist in the real world. They are not something we can design an experiment to measure, like we use experiments to measure the value of G, the gravitational constant. They can’t be measured, because they don’t exist.
Finally, I draw bgdwx’s attention to something he might have missed in Dyson’s story about Fermi:
Clearly, Fermi and Dyson did NOT share bgdwx’s crazy idea that every physical constant is actually a free parameter in disguise …
w.
Sure why not. What it is doing is solving for the virtual potential temperature and specific humidity in the PBL using inputs from a restart or initialization file using the PBL model developed by Cheng et al. 2002 which I’ll call CCH. Based on my cursory read of that model the code looks reasonable. The tcheck subrountine checks the temperature profile of the PBL for consistency with the ground temperature. If there is an inconsistency it morphs the CCH computed profile into one that is consistent with the ground temperature. You might be wondering how an inconsistency might arise. The CCH model could have deficiencies (the authors even say as much) or more to the point of the comment in the code it could be that the inputs given to ModelE were themselves unreasonable. In some configurations ModelE operates on a 1+ million cell grid mesh meaning that it is very likely that one of the inputs into the CCH model was in fact unreasonable which would cause an unreasonable output. Given the extent of the spatial domain I think it is impractical to halt the model run as it seems likely that at least one of the inputs would always be unreasonable. And when you see how the correction is applied it make sense. You could probably argue that there are better alternatives to a linear interpolation. But that would also come at the cost of increased computational complexity. Again, maybe it’s because I’m programmer that deals with issues like these on a daily basis. I don’t know. All I can say is that there is nothing offensive to me regarding the code. Would I want to do it differently? Yeah…probably.
It looks like Fermi and Dyson’s ideas are consistent with my own. And as I’ve said repeatedly I don’t think every physical constant is a free parameter. I’ve even given many examples of ones that are not, why they are not, and why you want to avoid them if possible.
Willis, you might get a better correlation between surface temperature and solar energy absorbed by examining only a one degree area of the earth’s surface on a monthly basis. Then take a random square to compare to see how the results hold.
I really dislike this averaging of annual temperatures (GAT) and insolulation values as it requires introducing fudge factors from biased sources. The lack of uniform temperature collection points and homogenezation techniques to make up for that lack requires introducing a bias. In addition, the gradual daily change in obliquity creating the seasonal change in insolulation would demonstrate the global consistency of this effect on all the other one degree squares. In short we would expect to see NH and SH squares act the same given the same insolulation value in their respective season regardless if it is over land or water.
From this, the AGW crowd would have a very difficult time of explaining away the solar connection to surface temperature.
As usual, things are more complex than they seem at first sight …
w.
Interesting, could you show just the absorption only?
Figure 4 above.
w.
Annual absorption is the same at the equator and the polar regions?
Are clouds and ice having the same effect?
Sunshine hours have definitely gone up here, notably since the AMO shifted to its warm phase from 1995:
Thanks, Ulric. I need to do a post on that question.
w.
Please do, I look forward to it.
Willis, great post, thank you.
From the dataplots we can see a lot of variation in time of the surface temperature and of the absorbed solar radiation, figure 1.
Also there are data points deviating from the trend line in figure 2 :
Temp = trend*Sol+m
If the atmospheric carbon dioxide concentration have any influence on the surface temperature then we can do an extended trend analysis like in :
Temp = trend_b*Sol+trend_c*CO2_Conc+m_b
What will the statistics tell about the two trend figures?
Kind Regards
Anders Rasmusson
Thanks, Anders.
The problem with that plan is that the change in CO2 forcing is basically a straight line. As a result, it will be assigned a large importance in any dataset with a trend.
Here, for example, is temperature as a function absorbed solar plus CO2 forcing, as well as temperature vs absorbed solar plus time of absorbed solar (a straight line). “allt2” is temperature, “absol” is absorbed solar, “coforce” is CO2 forcing.
And here is the same, but using the timeline in place of CO2 forcing
As you can see, almost no difference in significance (p-value, listed as “Pr(>|t|)” or R-squared. It’s a recurring problem with trying to assign causation to CO2 forcing.
w.
Ok, thanks a lot. Just to compare those statistic figures, it would be interesting to also see the corresponding statistics for the simple : Temp=trend*Sol+m
We got trend=0,22 °C/(W/m2) from your post. What about R-squared and the other statistics?
Kind Regards
Anders Rasmusson
Willis wrote: “my monkey mind started thinking about the relationship between how much sunshine is absorbed at the surface and the surface temperature. Here’s the CERES satellite data showing the relationship”
Well, Willis, maybe your “monkey mind” should run some of these ideas past a physicist before you post them, in order to look like slightly less of an idiot. Your “monkey mind” does not know what either “sunshine”, “temperature”, or “absorbed” means, so you trying to explain how these quantities are related at the surface by means of satellite measurements is not going to enlighten anyone about anything. While we are at it, where are the error bars?
You probably forgot that two gentlemen much smarter than you have already thought about these relationships, and they developed a law of physics that describes them. It is the Stefan-Boltzmann radiant heat transfer law, and the two gentlemen are of course Stefan and Boltzmann. Interestingly, what they found is that radiant energy transfer (and power is a measure of the rate of this transfer) depends on the temperatures of the two objects concerned, not the other way around.
So “°C per W/m2” is backwards. Just like your monkey mind…
Steve, were you born an arrogant ασσhole, or did you have to practice? Let me suggest that you dial back on the aggro. I suspect you’re a smart guy, but you’re sure doing your best to dispel that impression.
And why the anonymity? Got something to hide? I find that folks without the courage to sign their words are far more likely to indulge in your style of nasty personal attacks, because their words can never come back to haunt them or be recognized by people who actually know them … how about you stand up like a man and take off the mask?
To return to the issues in question, I gave the source of the data that I used. It’s at the bottom of the graphic. It’s the CERES EBAF 4.1 dataset. If you’d taken the trouble to look at it rather than jumping to attack me based on your misconceptions, you’d have seen that It includes both surface downwelling and surface reflected solar data. Here’s a list of the available datasets.
The difference between the CERES “Surface Shortwave Flux – Down” and the CERES “Surface Shortwave Flux – Up” datasets listed in the graphic above is the surface absorbed solar flux.
But then, you’d have known that if you’d done your homework.
If you think the CERES folks don’t understand what either “sunshine”, “temperature”, or “absorbed” means, I strongly suggest that you take it up with them …
As to the S-B Law, sure, I’ve used it for decades. Not sure what your point is there. If you’d quote what I said that you think contravenes the S-B law, please quote my exact words and I’m happy to discuss it.
w.
PS—Literally hundreds of physicists have read my work. When they’ve had objections, which naturally happens, we generally discuss them like gentlemen.
Let me invite you to do the same.
Willis, I am not anonymous. I cannot get WordPress (since the login system was implemented) to show my full name, no matter how I try. Others seem to have the same issue (not everyone, but a few people). But I do have a full name. It is Steve Keppel-Jones. You know, one of those pesky guys who keeps pointing out your constant physics errors? I am not hiding anything, because I take care to write only true physics statements, unlike you.
Remember, I tried to discuss your physics errors like a gentleman, but guess what I got in return? Logical fallacies, name-calling (“pond scum”, “pig”, “horse”, etc.), psychological projection, and finally, your favourite word, “Pass”. That is not how scientists discuss science, Willis. Are you sure I am the a$$h013 here?
I have tried to point out your errors many times. Many others have too. Every time, we get the same idiotic responses from you – “but the scientists said so”, “why would they lie”, “you’re wrong but I’m not going to tell you why”, “Pass! Pass! Pass!”, “Pig”, etc. You do not get to claim the moral or scientific high ground here, Willis. Especially for a guy who hasn’t studied physics in 50 years, at barely a first-year university level, spent all of the intervening time fishing, good for you, and whose information was apparently already 50 years out of date back then!
I would like to know which physicist agreed with your oft-repeated claim that SURFRAD scientists have measured positive downwelling longwave IR at the surface at night, many times, all over the world. I need to have a chat with whoever that guy is, because he’s wrong, just like you. And it is easy to show that directly, from first principles. You don’t know what those principles are, so naturally you can’t defend your own point, but if you know of a physicist who thinks he can, I’d like to talk to him.
I am not taking any of CERES’s errors up with the CERES folks because what they do is their business, and they are not paid to tell the truth any more than the SURFRAD folks are. They certainly won’t listen to me. But when you post their bad ideas and data here, uncritically, you are continuing to support folks involved in “one of the most arrogant, least scientifically supported naked power grabs in all of history”. The power grab you described as being run by people who “don’t have your best interests at heart” and therefore you shouldn’t “buy into it”. All of that is true. Yet here you are, once again buying wholesale into their nonsense and promoting it for them.
If you want to measure downwelling shortwave radiant energy at the surface, or surface temperature for that matter, why would you use a satellite? It’s not at the surface, is it? What physicist told you that this would be a good idea? Therefore “surface shortwave flux – down” and “surface shortwave flux – up” measured from a satellite are just guesses, and there’s a whole atmosphere in the way too, so they’re not going to be very good guesses at the best of times. They can try to “adjust” and “correct” for all those problems, but the more “adjustments” and “corrections” they add, the bigger the error bars get. Where are the error bars, Willis? I’m sure you know that error bars are a critical parameter of any physical science endeavour. But you (and apparently CERES) left them out. Why?
Of course, one of the things that real scientists (as opposed to you or the “climate scientists”) need to be able to do is to choose the correct instrumentation and experimental setup for the question at hand. A question about what’s happening at the surface is, of course, best answered by instruments at the surface. I recommend a pyranometer and a thermometer sitting on the ground, the same way SURFRAD is set up. Instead, you are trying to answer SURFRAD-type questions with a SATELLITERAD-type experiment. Bizarre… but not totally unexpected when a fisherman gets involved in physics, I suppose.
You claim that you know what the S-B law means, but you obviously don’t, otherwise you wouldn’t write “°C per W/m2”. (Indeed, your entire “Steel Greenhouse” farce shows that you have no idea what the S-B law says, or why it says it, never mind what “radiation”, “power”, or “energy” mean, or how they are related to each other. Remember when I asked you to define all those thermodynamics terms in your own words? You said “I’m not defining those words for you! Bugger off! Pass! Pig wrestling! Lead a horse to water but you can’t make him think!” – slightly paraphrased, but extremely unscientific, arrogant, and ignorant of you all the same. And the question still stands) Anyway, the S-B law says the relationship between temperature and radiant power transfer is the other way around from what you wrote – i.e. radiant power transfer depends on temperature. Meaning, W/m^2 per °C. Why not write these relationships correctly instead of backwards, the way the arrogant and scientifically unsupported “climate scientists” do it?
stevekj July 1, 2023 12:48 pm
Thank you. I’ve never heard of that happening, so I made an incorrect assumption. You have my sincere apology for me thinking you were deliberately hiding your name. Clearly, I was 100% wrong.
As to the rest? Sorry, not interested. Go bother someone else.
w.
Willis wrote: “Sorry, not interested.”
And, that would be why folks who know their physics call you an “idiot” instead of a “gentleman” or a “scholar”. You could learn some real physics, you know. I am quite sure of it. But definitely not with an attitude like that.
Nope. I’m not interested because you say ridiculous things like this:
“Anyway, the S-B law says the relationship between temperature and radiant power transfer is the other way around from what you wrote – i.e. radiant power transfer depends on temperature.”
You say that temperature does NOT depend on radiant power transfer. In other words, your claim is that an increase in radiant power transferred to the earth’s surface does NOT increase the temperature … in other words, you are saying that increasing sunshine doesn’t warm the earth.
The S-B law says:
Radiant power transfer = σ ε T^4
Where σ is the S-B constant, 5.67e-8, ε is emissivity, and T is the temperature in Kelvins.
So yes, you’re correct. Radiant power transfer does depend on temperature.
But we can solve the equation for T, which gives us
T = (Radiant power transfer / (σ ε))^1/4
So while radiant power transfer is indeed a function of temperature, temperature is also a function of radiant power transfer.
And I simply can’t deal with that level of ignorance.
For example, lots of scientists have looked at my “steel greenhouse” post. You’re one of the very few who seem to misunderstand it and claim that it violates thermodynamic laws.
For folks not familiar with my work on the question, see my posts “The Steel Greenhouse” and “People Living In Glass Planets“.
If you (or stevekj) think you’ve found an error in either of those, please quote my exact words and show (demonstrate, not simply claim) why my words are incorrect.
w.
Oh and another thing occurred to me. When you wrote “PS—Literally hundreds of physicists have read my work. When they’ve had objections, which naturally happens, we generally discuss them like gentlemen.”, did you mean to write “hundreds of climate scientists“?
Because no actual physicist who has passed an actual physics exam would sign off on anything you have written about thermodynamics, or quantum mechanics for that matter – it is all wrong. Every word of it that I’ve seen, in several different forums, for many years. And you are completely impervious to being corrected, by me or anyone else who knows better – hence my slightly acerbic tone. There is no point in being gentlemanly with someone who is as arrogantly ignorant and closed-minded as you are. I tried. Many others did too. And I am very polite with people who are willing to learn, like Ferdinand Engelbeen for instance. I corrected his misconceptions about thermodynamics politely, over the course of many days. Eventually he saw that his previous view (the same as yours) was false, because it led to an obvious contradiction. You’re not smart enough to get that far, sadly. And certainly not for lack of trying on my part, heaven knows.
(To be fair, you wrote one correct thing once, when you said “electromagnetic radiation is a form of energy”. But you only wrote that once, that I saw, and nothing else you wrote is consistent with that statement – and when we pressed you to explore the consequences of this statement on the rest of your false physics, you naturally refused, perhaps hazily glimpsing a glaring contradiction off in the distance waiting to horribly embarrass you. So I am fairly sure you only got that one right by accident.)
But, to return to my main point here, climate scientists are not physicists, Willis. The field of “climate science” bears approximately the same relationship to physics as astrology bears to astronomy. They use the same words, yes, but they use them in completely different (and generally wrong) ways. And even that is probably being unfair to astrologers.
This may seem, on the surface, astonishing to a fisherman or a cowboy, who is not used to thinking that all the fish or all the cows are somehow lying en masse, but remember, climate scientists are not in the business of selling the truth, in the way that, say, a university physics professor is. They are in a different business. They are in the business of selling “one of the most arrogant, least scientifically supported naked power grabs in all of history” (those are your words, in case you forgot already). There is no place for truth in that endeavour. None. In fact the truth would be highly counterproductive for them. That’s why WUWT exists at all. And because they are government employees, no one in any position of authority is going to grade them on their work. Except us here on the internet. Well, not you, obviously. But some of the rest of us who do know our physics, anyway.
So someone around here who is more creative than me coined the phrase “climate fiziks” (or some similar variation) to refer to what they are doing. It is not physics. Do not confuse the two. And please do not insult actual physicists by referring to “climate scientists” as “physicists”.
Steve, you’re free to keep believing that if it helps you to sleep.
Me? Not interested. I’d strongly advise you to try out your line of patter on someone else. It doesn’t work with me.
w.
“It doesn’t work with me.”
I know that, Willis. I’ve tried to educate you before, as have many others, with the same result. They do call it insanity when you do the same thing over and over and keep getting the same result. And maybe trying to teach you physics is insane. But in the same way that you come from a line of fishermen, I come from a line of professors, so it’s in the blood, and I have to keep trying. It’s just what I do…
“Tried to educate me”?
You are truly hilarious.
w.