I’ve been looking at the Nikolov and Zeller paper again. Among other things, they claim to be able to calculate the surface temperature Ts of eight different planets and moons from knowing nothing more than the solar irradiation So and the surface pressure Ps for each heavenly body. Dr. Zeller refers to this as their MIRACLE equation. He says:
Why aren’t you all trying to disprove our MIRACLE equation rather than banging your heads against walls trying to prove or disprove who knows what and exclaiming you have problems with this or that? The question is how can we possibly have done it – there is no question that our equations work – if you haven’t verified that it works, why haven’t you? […] Why aren’t you thinking: “hmmmm, N&Z have given us an equation that lo-and-behold when we plug in the measured pressures and calculate Tgb as they suggest, gives us a calculated Ts that also matches measured values! You can’t disprove the equation? So maybe we are cooking the data books somehow, but how?
This is supposed to be evidence that their theory is correct, and people keep telling me ‘but they’ve got real evidence, they can make predictions of planetary temperatures, check it out”. Plus it’s hard to ignore an invitation like Dr. Zellers, so I checked it out.
Figure 1. These are not the equations you are looking for.
They first postulate something called the “Near-surface Atmospheric Thermal Enhancement” or “ATE” effect that makes the earth warmer than it would be without an atmosphere.
The “ATE effect” is measured by something called Nte(Ps), which is defined and estimated in their paper as follows.
where Nte(Ps) is a measure of the “Near-surface Atmospheric Thermal Enhancement” effect.
Nte(Ps) is defined as the actual average surface air temperature of the planet Ts divided by the theoretical “graybody” temperature of the planet Tgb calculated from the total solar insolation So of the planet. Nte(Ps) is estimated using a fitted function of the surface pressure of the planet Ps.
Let me simplify things a bit. Symbolically, the right part of equation (7) can be written as
Nte(Ps) = e^(t1 * Ps ^ t2 + t3 * Ps ^ t4) (7Sym)
where “e” is the base of natural logs and Ps is the surface pressure on the planet or moon. There are four tunable parameters (t1 through t4) that are “fitted” or tuned to the data. In other words, those values are repeatedly adjusted and tuned until the desired fit is obtained. This fitting can be easily done in Excel using the “Solve…” menu item. As you’d expect with four parameters and only eight datapoints, the fit is quite good, and their estimate is quite close to the actual value of Nte(Ps).
Amusingly, the result of equation (7) is then used in another fitted (tuned) equation, number (8). This is:
where So is total solar irradiation.
This is their piece de resistance, their MIRACLE equation, wherein they are saying the surface temperature of eight different planets and moons can be calculated from just two variables— Pr, the surface pressure, and So, the total Solar irradiation. This is what amazes the folks in the crowd so much that they write and tell me there is “evidence” that N&Z are right.
Obviously, there is another tuned parameter in equation (8), so we can rewrite this one symbolically as:
Ts = t5 * (Solar + adjustment ) ^ 1/4 * Nte(Ps). (8Sym)
Let me pause a minute and point something out about equation (8). The total solar irradiation Solar ranges from over 9,000 W/m2 for Mercury down to 1.51 W/m2 for Triton. Look at equation 8. How will adding the adjustment = 0.0001325 to any of those values before taking the fourth root make the slightest bit of difference in the result? That’s just bizarre, that is. They say they put it in so that the formula will be accurate when there is no solar, so it will give the background radiation of 3 Kelvins. Who cares? Truly, it changes Ts by a maximum of a thousandth of a degree for Triton. So for the moment let me remove it, as it makes no practical difference and it’s just confusing things.
Back to the tale. Removing the adjustment and substituting equation 7 into equation 8 we get:
Ts = t5 * Solar^0.25 * e^(t1 * Ps ^ t2 + t3 * Ps ^ t4) (eqn 9)
This is amazing. These guys are seriously claiming that with only eight datapoints and no less than five tunable parameters , they can calculate the surface temperature of the eight planets knowing only their surface pressure and solar irradiation. And with that many knobs to turn, I am sure they can do that. I did it on my own spreadsheet using their figures. I get about the same values for t1 through t5. But that proves nothing at all.
I mean … I can only stand in awe at the sheer effrontery of that claim. They are using only eight datapoints and five tunable parameters with a specially-designed ad-hoc equation with no physical basis. And they don’t think that’s odd in the slightest.
I will return to this question of the number of parameters in a bit, because even though it’s gobsmacking what they’ve done there, it’s not the best part of the story. Here’s the sting in the tale. We can also substitute equation (7) into equation (8) in a slightly different way, using the middle term in equation 7. This yields:
Ts = t5 * Solar^0.25 * Ts / Tgb (eqn 10)
This means that if we start out by knowing the surface temperature Ts on the right side of the equation, we can then calculate Ts on the left side … shocking, I know, who would have guessed. Let’s check the rest of the math in equation (10) to see why that works out.
Upon inspection it can be seen that the first part of the right side of equation (10),
t5 * Solar^0.25
is an alternate form of the familiar Stefan-Boltzmann equation relating temperature and radiation. The S-B equation can be written as
T = (Solar / c1) ^ 0.25.
where T is temperature and c1 is a constant equal to the S-B constant times the emissivity. We can rewrite this as
T = 1/(c1^0.25) * Solar^0.25
Setting another constant c2 equal to 1 / (c1^0.25) gives me the Stefan-Boltzmann equation as:
T = c2 * Solar^0.25
But this is exactly the form of the first part of the right side of equation 10. More to the point, it is an approximation of the graybody temperature of the planet Tgb.
We can check this by observing that if emissivity is .9 then constant c1 is 5.103E-8, and c2 is therefore about 66. However, that value will be reduced by the rotation of the planet. Per the N&Z formula in their latest post, that gives a value of about 27.
Their fitted value is 25, not far from the actual value. So curiously, what it turns out they’ve done is to estimate the Stefan-Boltzmann constant by a bizarre curve fitting method. And they did a decent job of that. Actually, pretty impressive considering the number of steps and parameters involved.
But since t5 * Solar^0.25 is an estimation of the graybody temperature of the planet Tgb, that means that Equation 10 reduces from
Ts = t5 * Solar^0.25 * Ts / Tgb (eqn 10)
to
Ts = Tgb * Ts / Tgb.
and finally to
Ts = Ts
TA-DA!
CONCLUSION
Let me recap the underlying effect of what they have done. They are looking at eight planets and moons.
1. They have used an equation
e^(t1 * Ps ^ t2 + t3 * Ps ^ t4)
with four free parameters to yield an estimate of Ts/Tgb based on surface pressure. As one would expect given the fact that there are half as many free parameters as there are data points, and that they are given free choice to pick any form for their equation without limit, this presents no problem at all, and can be done with virtually any dataset.
2. They have used an equation
t5 * Solar^0.25
with one free parameter in order to put together an estimate of Tgb based on total planetary insolation. Since Tgb does depend inter alia on planetary insolation, again this presents no problem.
3. They have multiplied the two estimates together. Since the result is an estimate of Tgb times an estimate of Ts/Tgb, of course this has the effect of cancelling out Tgb.
4. They note that what remains is Ts, and they declare a MIRACLE.
Look, guys … predicting Ts when you start out with Ts? Not all that hard, and with five free parameters and a choice of any equation no matter how non-physically based, that is no MIRACLE of any kind, just another case of rampant curve fitting …
Finally, there is a famous story in science about this kind of pseudo-scientific use of parameters and equations, told by Freeman Dyson:
We began by calculating meson–proton scattering, using a theory of the strong forces known as pseudoscalar meson theory. By the spring of 1953, after heroic efforts, we had plotted theoretical graphs of meson–proton scattering. We joyfully observed that our calculated numbers agreed pretty well with Fermi’s measured numbers. So I made an appointment to meet with Fermi and show him our results. Proudly, I rode the Greyhound bus from Ithaca to Chicago with a package of our theoretical graphs to show to Fermi.
When I arrived in Fermi’s office, I handed the graphs to Fermi, but he hardly glanced at them. He invited me to sit down, and asked me in a friendly way about the health of my wife and our newborn baby son, now fifty years old. Then he delivered his verdict in a quiet, even voice. “There are two ways of doing calculations in theoretical physics”, he said. “One way, and this is the way I prefer, is to have a clear physical picture of the process that you are calculating. The other way is to have a precise and self-consistent mathematical formalism. You have neither.
I was slightly stunned, but ventured to ask him why he did not consider the pseudoscalar meson theory to be a selfconsistent mathematical formalism. He replied, “Quantum electrodynamics is a good theory because the forces are weak, and when the formalism is ambiguous we have a clear physical picture to guide us. With the pseudoscalar meson theory there is no physical picture, and the forces are so strong that nothing converges. To reach your calculated results, you had to introduce arbitrary cut-off procedures that are not based either on solid physics or on solid mathematics.”
In desperation I asked Fermi whether he was not impressed by the agreement between our calculated numbers and his measured numbers. He replied, “How many arbitrary parameters did you use for your calculations?”
I thought for a moment about our cut-off procedures and said, “Four.”
He said, “I remember my friend Johnny von Neumann used to say, with four parameters I can fit an elephant, and with five I can make him wiggle his trunk.” With that, the conversation was over. I thanked Fermi for his time and trouble, and sadly took the next bus back to Ithaca to tell the bad news to the students.
The Nikolov and Zeller equation contains five parameters and only eight data points. I rest my case that it is not a MIRACLE that they can make the elephant wiggle his trunk, but an expected and trivial result of their faulty procedures.
My regards to everyone,
w.
PS—There is, of course, a technical term for what they have done, as there are no new mistakes under the sun. It is called “overfitting”. As Wikipedia says, “Overfitting generally occurs when a model is excessively complex, such as having too many parameters relative to the number of observations.” Five parameters is far, far too many relative to eight observations, that is a guaranteed overfit.
PPS—One problem with N&Z’s MIRACLE equation is that they have not statistically tested it in any way.
One way to see if their fit is even remotely valid is to leave out some of the datapoints and fit it again. Of course with only eight datapoints to start with, this is problematic … but in any case if the fitted parameters come out radically different when you do that, this casts a lot of doubt on your fit. I encourage N&Z to do this and report back on their results. I’d do it, but they don’t believe me, so what’s the point?
Aother way to check their fit is to divide the dataset in half, do the fit on one half, and then check the results on the other half. This is because fitted equations like they are using are known to perform very poorly “out of sample”, that is to say on data not used to fit the parameters. Given only eight data points and four parameters for equation 7, of course this is again problematic, since if you divide the set in half you end up with as many parameters as data points … you’d think that might be a clue that the procedure is sketchy but what do I know, I was born yesterday. In any case I encourage N&Z to perform that test as well. My results from that test say that their fit is meaningless, but perhaps their test results will be different.
[UPDATE] One of the commenters below said:
Willis – go ahead – fit an elephant. Please!
Seriously N&Z are only demonstrating in algebra what has been observed in experiments, that heating a gas in a sealed container increases both pressure and temperature.
OK, here’s my shot at emulating the surface temperature using nothing but the data in the N&Z chart of planetary body properties:
Figure 1. Willis’s emulation of the surface temperature of the planetary bodies.
My equation contains one more variable and two less parameters than the N&Z equation. Remember their equation was:
Ts = 25.3966 * Solar^0.25 * e^(0.233001 * Pressure ^ 0.0651203 + 0.0015393 * Pressure ^ 0.385232)
My equation, on the other hand, is:
Ts = 0.8 * Tgb + 6.9 * Density + 0.2 * Gravity)
Note that I am absolutely not making any claim that temperature is determined by density and gravity. I am merely showing that fitting a few points with a few variables and a few parameters is not all that difficult. It also shows that one can get the answer without using surface pressure at all. Finally, it shows that neither my emulation nor N&Z’s emulation of the planetary temperatures are worth a bucket of warm spit …
[UPDATE 2] I figured that since I was doing miracles with the N&Z miracle equation, I shouldn’t stop there. I should see if I could beat them at their own game, and make a simpler miracle. Once again, their equation:
Ts = 25.3966 * Solar^0.25 * e^(0.233001 * Pressure ^ 0.0651203 + 0.0015393 * Pressure ^ 0.385232)
My simplified version of their equation looks like this:
Ts = 25.394 * Solar^0.25 * e^(0.092 * Pressure ^ 0.17)
Curiously, my simplified version actually has a slightly lower RMS error than the N&Z version, so I did indeed beat them at their own game. My equation is not only simpler, it is more accurate. They’re free to use my simplified miracle equation, no royalties necessary. Here are the fits:
Figure 2. A simpler version of the N&Z equation 8
Again, I make no claim that this improves things. The mere fact that I can do it with two less tuned parameters (three instead of five) than N&Z used does not suddenly mean that it is not overfitted.
Both the simplified and the complex version of the N&Z equations are nothing but curve fitting. This is proven by the fact that we already have three simple and very different equations that hindcast the planetary temperatures. That’s the beauty of a fitted equation, if you are clever you can fit a lot using only a little … but THAT DOESN’T MEAN THAT PRESSURE DETERMINES TEMPERATURE.
For example, I can do the same thing without using pressure at all, but using density instead. Here’s that equation:
Ts = 25.491 * Solar^0.25 * e^(0.603 * Density ^ 0.201)
And here’s the results:
Figure 3. An emulation of the planetary temperatures, using density instead of pressure.
Does this now mean that the planetary temperature is really controlled by density? Of course not, this whole thing is an exercise in curve fitting.
w.
Thomas L says:
January 30, 2012 at 12:41 am
I’m not impressed with any theory that claims an accuracy of a hundredth of a degree C for the global temperature. We don’t know global temperature to the nearest hundredth of a degree, either in theory or practice.
w.
John Day says:
January 30, 2012 at 2:56 am
The weakest part of ATE, IMHO, is this Hoelder integration stuff, which ends up increasing the GHE enhancement from 33K to 133K. Do you accept that? If you can show that’s not true and that 33K is the actual enhancement, then the N&Z collapses internally, based on its own principles.
It’s certainly not true as it depends on the assumption of zero heat capacity for the planet’s surface, which is nowhere near true, even for the Moon. Rather than a virtually zero surface temperature the NASA data shows around 100K (N&Z show this in their post). N&Z have studiously avoided addressing this, despite Anthony prompting them too!
John Day says:
Yes…but the atmosphere is above the solid/liquid surface. That is the important thing. It turns out that an atmosphere that absorbs radiation can make the planet’s surface warmer than would be possible if the atmosphere did not absorb any of the radiation emitted by the surface.
This makes no sense. You seem to want ATE to serve as the step “…And then a miracle occurs…”. Saying that if ATE were true then it would explain the warmth is TRULY begging the question. You have to be able to explain how a method not involving the absorption of radiation emitted by the planet’s surface can still lead to a surface temperature that exceeds the limit that seems to be placed on it by the constraint of radiative balance.
That is, new mechanisms to explain the warmth still have to obey known physical laws…Or, alternatively, if they don’t, then extraordinary evidence needs to be presented to confirm that these physical laws really are violated.
I have already explained this in great detail and it is unclear to me what part of it you don’t understand. It is true that a planet that has absolutely no atmosphere (and no liquid on the surface and such a slow rotation rate that any heat capacity can be neglected) would have an average temperature about 133 K colder. However, as I have explained many times, we already understand how to get rid of ~100 K of that 133 K discrepancy: It is a simply artifact of the fact that there are lots of different temperature distributions having different average temperatures that all have the same total emission of radiation from the surface and hence all are equally plausible solutions to the condition of radiative balance for the planet. Which distribution (and hence which average temperature) occurs depends on the issues that I talked about: distribution of insolation, heat transport on the planet, rotation rate of the planet, heat capacity, …
However, the highest average temperature a blackbody can have and still be emitting an average of 240 W/m^2 is 255 K. That average temperature occurs for a perfectly uniform distribution. And, it is the difference between 255 K and our planet’s actually average surface temperature of about 288 K (with corresponding average emission of 390 W/m^2 or so) that cannot be explained by any other way except that the atmosphere absorbs some of the radiation emitted by the surface.
Again, it puzzles me why these basic facts seem to be so difficult for you and some others to get your minds around. If you could tell me where you lose the flow of the logic, it would help.
Willis Eschenbach says:
January 30, 2012 at 12:08 am
As an aside, here’s one problem I have with Venus:
As can be seen from above link, partial pressure of CO2 in the Venusian ground atmosphere amounts to ~89 bar and ground temperature to 737 K on average. The “critical point” data of CO2: p(crit.) = 72.9 bar; T(crit.) = 304 K (handbook of physics).
A substance beyond the critical point is neither gas nor fluid, though in general referred to as supercritical fluid. It surely is not a gas and very surely not a trace gas. It is very important to note that the evaporation enthalpy of a substance beyond the critical becomes zero; this means there exists no cooling effect by evaporation, which is a very important energy transfer and cooling mechanism on earth.
There’s a more fundamental problem with the N&Z fitting using the Venus data arising from the supercritical CO2 lower atmosphere. The Ideal Gas Law which they use in their calculation doesn’t apply! For an Ideal gas PV/nRT=1.0, at 90 atm CO2 and 273K PV/nRT =~0.3, for example. To apply their model they’d need to take account of the Compressibility factor at the surface conditions of the Venusian atmosphere. The high temperature reduces the compressibility effect but I see no evidence they’ve even considered it.
@Phil.
> It’s certainly not true as it depends on the assumption
> of zero heat capacity for the planet’s surface, which is
> nowhere near true, even for the Moon.
Let me see if I understand correctly. Are you claiming that N&Z stated explicitly that there is zero heat capacity, or that they didn’t include a heat-capacity term in their formulation?
If the latter, giving them the benefit of doubt, perhaps it is the case that heat capacity is not needed here.
Yes, it’s obvious that all solid materials have non-zero heat capacity. But surprisingly (to me at least), at the planetary level, the surface temperature estimate returned by the Stefan-Boltzmann equation seems to be independent of surface composition (and therefore surface heat capacity).
This is probably, as Joel pointed out, due to the broad, continuous spectra of solids and liquids. At a planetary scale they all tend to look alike, so you only need to supply insolation and emissivity to get a good wag on the surface temperature.
Does that help resolve the heat-capacity problem you’re talking about?
😐
@joel Shore
> This makes no sense. You seem to want ATE to serve as the step
> “…And then a miracle occurs…”. Saying that if ATE were true then
> it would explain the warmth is TRULY begging the question.
No, I’m not asking you to believe in miracles, but merely to accept the premise for the sake of argument, because we have not been fully informed on how the ATE step actually works i.e. we’re still waiting for N&Z to publish enough details to get our heads completely around it. At that point we might discover that it specifies a perpetual-motion machine, or similar nonsense. But I don’t think we can confidently state we understand ATE to make that claim (correct me if I’m wrong).
At stake here is credibility. As soon as it is learned that a theory actually violates energy, the theory is toast. I wouldn’t give it the time of day. So, I don’t think it’s fair to give the theory that kind of mortal handicap until we understand it completely under the terms of their arguments.
So, we can’t just say “I don’t completely know how your ATE thingy works, but it obviously violates energy conservation because _we_ believe GHG’s are the only true way for this to happen, even though you claim it is a different (and the correct) mechanism to explain this warming.
Thus basing your falsification merely on the fact that it doesn’t recognize GHG-GHE would be begging the question. Or maybe it would help to call it begging the falsification, since falsification is the tool you have chosen to dislodge this question, in a formal logic sense.
Now, on the other hand, if you think you understand ATE enough to prove it violates energy then by all means call them out on it. (But, again, you can’t say it’s because they don’t believe in GHG’s, because that’s theory _they_ are trying to dislodge .
As an illustration [git!], it would be like the followers of Ptolemy insisting that Copernicus was wrong because his theory didn’t have enough epicycles in it [or too many epicycles, according to the anonymous git]
😐
@joel Shore
> Again, it puzzles me why these basic facts seem to be so
> difficult for you and some others to get your minds around.
In my case, it’s partly because I was only “skimming” the N&Z issues then. So I missed some of your previous arguments. Sorry.
Let me look back and see I can get back up to date on the Hoelder issue.
Also, I have been studying your “basic stuff you need to know” above. I have a question or two about that, which I’ll also present to you for clarification later this evening. Thanks.
😐
[Phil, you accidentally copied and pasted into your comment about half the comments in the thread. Your comment ended up hundreds and hundreds of lines long. I couldn’t figure out which of the thousands of words were yours, so I’m sorry, but you’ll have to post it again. My apologies, but it nearly doubled the length of the comments thread. w.]
John Day says:
January 30, 2012 at 2:08 pm
Sorry, John, but science doesn’t work that way. Scientists don’t accept outlandish claims “for the sake of argument”.
The problem is exactly what you point out. Nobody can explain, and therefore nobody understands, N&Z. Until we do understand it, I for one will never accept it “for the sake of argument”. Why on earth would I want to do that? How can I argue either in favor of or against something I don’t understand?
Until you can explain it, it’s a claim of a miracle. My policy is … never discuss miracles with a true believer … or with an unbeliever, for that matter. Call me crazy, but I’m into facts.
w.
John Day:
I think Willis very well-expressed what I would say in response to you. And, of course, when N&Z publish more details, I will read them. But, I think at this point, believing that they are going to come up with some sort of compelling argument about how their “theory” can satisfy conservation of energy is about as likely as believing that the next time that I play basketball, I am going to play as well as Michael Jordan. One cannot, strictly speaking, rule out the possibility…but nothing so far has given us any reason to expect this.
We already know that what we have gotten from N&Z are boneheaded errors: They don’t know how to add in convection to a radiative model of the greenhouse effect (and think they are the only ones who do know how to do it). They don’t understand how to apply conservation of energy to a planet that is receiving energy from the sun. They believe that a 4-free-parameter fit constitutes a miracle.
Willis:
“In other words, what’s happening on Venus is unlikely to be directly comparable to what’s happening on the Earth. I doubt we can draw many inferences that are common to both worlds. So I find all discussion of Venus in this context to be a total diversion. We don’t understand Earth’s climate, and our understanding of Venus’s climate is orders of magnitude worse.
Now, I’d said that until you could give me the elevator speech on Huffman, I wouldn’t touch his ideas about Venus at all. I repeat that again in spades. His writing is impenetrable to me. He uses vague words in a vague way, and tries to explain physics with metaphors rather than math. If you can explain his theory with an elevator speech, fine. Otherwise …”
OK, I’m done, Willis. Especially if even Huffman can’t even come to his own defense! I thought he had presented some good empirical evidence. Perhaps not.
BTW, FWIW, I was not in any way trying to set a standard to which YOU would aspire, but one which I could aspire.
And the “messing” is arrogant, to say the least.
Phil. says:
January 30, 2012 at 2:58 pm
[Phil, you accidentally copied and pasted into your comment about half the comments in the thread. Your comment ended up hundreds and hundreds of lines long. I couldn’t figure out which of the thousands of words were yours, so I’m sorry, but you’ll have to post it again. My apologies, but it nearly doubled the length of the comments thread. w.]
Sorry, using iPhone I didn’t see that!
@joel Shore
> One cannot, strictly speaking, rule out the possibility…
> but nothing so far has given us any reason to expect this.
That is essentially the attitude that I am advocating. Possibly correct until proven false.
With respect to that 100K Hoelder discrepancy: earlier you said …
So you agree with Nikolov on the method of integration, but you interpret the 100K difference has being caused by the non-uniform distribution of surface temps, so Tmean < Tmax. Else if strictly uniform then Tmean=Tmax. That actually makes sense now.
What are the major non-uniformities? Perhaps the night-vs-dayside and equator-vs-polar distributions?
John Day says:
January 30, 2012 at 1:35 pm
@Phil.
> It’s certainly not true as it depends on the assumption
> of zero heat capacity for the planet’s surface, which is
> nowhere near true, even for the Moon.
Let me see if I understand correctly. Are you claiming that N&Z stated explicitly that there is zero heat capacity, or that they didn’t include a heat-capacity term in their formulation?
Their basic assumption that the surface temperature is zero where the insolation is zero implicitly means zero heat capacity. They include no heat capacity term in their formula.
If the latter, giving them the benefit of doubt, perhaps it is the case that heat capacity is not needed here.
No benefit of the doubt it’s flat out wrong, as is clearly shown by the NASA data for the moon which they quoted.
Does that help resolve the heat-capacity problem you’re talking about?
😐
No, it’s responsible for their large error in the calculation of the surface temperature because of the large effect on the temperature distribution.
jae says:
January 30, 2012 at 7:44 pm
My apologies, jae, if I did not understand what you meant when you said:
Trying to dictate to a man the conditions under which you will concede that he has won is … well, let me just say that I’m not the only man who might misinterpret your statement.
You don’t get to say something like ‘you haven’t won until you do a back flip holding a coffee cup’. You also don’t get to say I haven’t won because I haven’t read some jerk you think has the inside track on perpetual motion. And yes, I am sensitive to that, perhaps overly so.
But it’s because I get that kind of garbage all the time ‘but willis, you must read this garbage, you must consider this impossibility’.
So I do apologize, jae, and I also want you to understand what your demand, that I consider your pet theory, looks like from this side.
My best regards to you,
w.
kuhnkat says:
January 30, 2012 at 9:38 pm
Although I appreciate your link, kuhnkat, I know about that stuff, there’s even a word for it. Serendipity, the lucky accident.
As I said before, the fact that people have stumbled over things in the dark doesn’t mean we should all go stumbling around in the dark, experimenting at random and hoping for the best.
So yes, Konrad may indeed be able to pull an experiment out of his … chance imaginings and have it provide valuable information. But that’s not how smart scientists play the experiment game, kuhnkat.
Smart scientists first study the theory until they understand all its parts. Then and only then, they design an experiment to either falsify or support whatever they see as the most vulnerable part of the theory. They are careful to design it to avoid as many confounding variables as possible, and to be able to measure the ones they can’t avoid.
Next, smart scientists say in advance what it is that they expect their results to be if the theory is right, and what they expect to find if the theory is wrong. That’s what the experiment is for, after all—to test the theory. The only way to do that is to design an experiment that can discriminate, that will give different results depending on whether it is true or false.
So sure, Konrad might find something stumbling around in the dark. It’s just that the smart money’s not betting on it. How will you interpret the results if you don’t understand the theory? How will you recognize the unusual if you don’t know what to expect?
w.
Willis,
you played cards and never figured out overplaying your hand away from the table??
John Day says [emphasis mine]:
January 30, 2012 at 8:13 pm
John, the attitude that Joel is advocating is not what you claim.
He is advocating the attitude of “Almost certainly incorrect until proven false” whenever dealing with folks who make claims contrary to the Laws of Thermodynamics.
As am I. They don’t call it a Law for nothing. It is a law because it applies everywhere in all situations and nobody has ever found a single instance of it being violated. Not one. So if someone says they have, their claims are almost certainly incorrect.
w
@Willis
> … the attitude that Joel is advocating is not what you _claim_.
No, I _claimed_ (after reading that comment by Joel) “That is essentially the attitude that I am advocating. ”
Putting words in my mouth again you are.
Everyone is entitled to advocate whatever floats their boat.
There is, in fact, a database for high temperature CO2 absorption/emission lines. It’s called HITEMP ( see description here). I calculated some spectra for a Venus-like atmosphere using HITEMP at Spectralcalc. A supercritical fluid is still more gas than fluid so it doesn’t behave quite like a black body. There are a few windows in the CO2 spectrum even at 700K and 92bar. However, they are mostly filled by water vapor and sulfur dioxide. As I remember, I calculated that at the surface, there was about 5 W/m² less downwelling radiation than upwelling radiation (out of ~14,000 W/m²). But the average flux from sunlight at the surface is at least 9 W/m².
@DeWitt Payne
> I calculated some spectra for a Venus-like
> atmosphere using HITEMP at Spectralcalc.
Very interesting. Is there a way to see some of the plots you generated?
John Day says:
Yes…In particular, the calculation that Nikolov and Zeller have done to determine the average temperature for an airless planet, T_sb, assumes that the temperature at a particular point on the planet’s surface at a particular time is just the value necessary so that this point is emitting back out into space exactly as much power as it is receiving from the sun. So, for example, the half of the planet that is not facing the sun is assumed to be at a temperature of absolute zero (which they then correct to ~3 K) because it isn’t receiving any power from the sun. These assumptions result in a very uneven temperature distribution on the planet and that is why their T_sb value is so low.
@joel Shore
> These assumptions result in a very uneven temperature
> distribution on the planet and that is why their T_sb value is so low.
Ok, I think I am understanding this, but that last sentence threw me a bit. If their distributions are very uneven, wouldn’t we expect T_sb to be low? i.e. lower than for a uniform distribution, given the same insolation? Is it just that 3K is too low? If so, what value is expected for the backside?
John Day,
Here’s the only one I could find quickly. It’s a plot of the absorption spectrum rather than emission, but it shows the windows with no other gas present than CO2. In case you can’t read the fine print, it’s a plot from 500-5500cm-1 with a path length of 500m, total pressure 92bar of which 96% is CO2. To get emission, one would multiply the Planck equation at each wavenumber and T=730K by the absorptivity at that wavenumber. If I did my sums correctly, peak emission at 730K is at 1431cm-1. I did the calculations in Excel and even broken down into smaller ranges, things were unwieldy. The spreadsheets got so large they wouldn’t open in XP, but would, barely, in Win7. The problem with line-by-line programs is that the resolution must be so high that an enormous number of points are generated. The text files were multi-megabytes.
John Day says:
January 31, 2012 at 12:32 pm
@joel Shore
> These assumptions result in a very uneven temperature
> distribution on the planet and that is why their T_sb value is so low.
Ok, I think I am understanding this, but that last sentence threw me a bit. If their distributions are very uneven, wouldn’t we expect T_sb to be low? i.e. lower than for a uniform distribution, given the same insolation? Is it just that 3K is too low? If so, what value is expected for the backside?
For the moon you’d expect more like 100K (they plot it from NASA in their response), that’s the result of not including surface heat capacty.
http://wattsupwiththat.files.wordpress.com/2012/01/image38.png
@Phil.
> For the moon you’d expect more like 100K (they plot it from
> NASA in their response), that’s the result of not including surface heat capacity.
So, I missed the previous message traffic on this. are you saying that N&Z have never responded at all to this issue? Seems rather crucial (now that I understand where you’re coming from, wrt heat capacity. Sorry for the confusion). I guess a faster rotating airless planet would have an even warmer backside.