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.
kuhnkat says:
We know that they did curve fitting because they told us that they did curve fitting ( http://wattsupwiththat.com/2011/12/29/unified-theory-of-climate/ ):
And, now they have admitted (or, strangely enough, bragged) that they shopped around extensively for the nonlinear function that they chose to fit with ( http://tallbloke.wordpress.com/2012/02/09/nikolov-zeller-reply-eschenbach/ ):
So, it takes a minimum of 4 parameters to fit 8 pieces of data…and, as Robert Brown has argued, of those 8, only 5 are fit well. (It is hard to notice that the other 3 don’t fit all that well because they are scrunched up on the left-hand side of the plot where N_TE vs pressure rises so steeply.)
That is not the sort of “miracle” that is worth abandoning conservation of energy and one’s senses to embrace!
Joel Shore,
you claim they were curve fitting. So, the parameters they are using have nothing to do with gravity, pressure, mass, albedo… Just a curve fitting exercise.
HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA
I just can’t resist pointing out this comment over at Tallbloke’s http://tallbloke.wordpress.com/2012/02/09/nikolov-zeller-reply-eschenbach/#comment-16985 wherein David M Hoffer has claimed to have found something that would be a violation of Holder’s Inequality!
It rests on the fact that David doesn’t realize that if you have 24 values, half of which are zero, you still have to divide by 24 and not 12 to get the average! (Actually, he did realize this when he computed the average of the power radiated, just not the temperature, so maybe it was just a simple arithmetic error on his part.)
Gotta love how these guys think they can find examples that violate a rigorously-proven mathematical theorem!
Willis Eschenbach, Re your February 10, 2012 at 10:59 am comment:
I’m going to let it play out. I haven’t even looked into what they’ve done in great depth yet. Beginning that is one of the things on the agenda for late tonight, for my curiosity. It’s an interesting, simple idea. It’s something that makes intuitive sense, but so do a lot of things including the mainstream AGW theory, which I don’t believe fits the data.
I do notice you seem to have an emotional investment in this over and above what the numbers are. I guess that makes sense; it’s your work and your competence that’s been called into question and vice versa.
But it’s not persuasive to your readers.
Anyway, persuading me should only be of minor concern as I’m one voice in a much larger debate, which will not ultimately be settled by popular vote (one hopes).
I also find Henrik Svensmark’s work interesting and am curious what you think of his model. There seems to be a good data fit on short and long timescales as well as a simple, intuitively easily to understand and visualize physical model of how it works.
Also, as far as notation goes, of course your notation is accepted. I just meant it would be that easier to visualize if created in Latex with raised superscripts, two lines, and the like. It’s an aesthetic point, not a substantive one. It was just a suggestion, and I know it’s a minor one.
But I kind of like mathematical symbolic aesthetics.
Keep up the work. I appreciate your efforts and thought on these matters.
I DO think that discounting someone’s future work in advance because they don’t immediately back down when challenged isn’t in keeping with the best principles of scientific inquiry. Even if they’re wrong, you’re not as prominent now as was Enrico Fermi then, and it may take them longer to come to the realization.
We are all psychological beings, after all. If that wasn’t the case, the climate debate would be going much smoother, wouldn’t it?
Since I last commented and before reading this comment, I’ve certainly been thinking about the conversation of energy and Anthony’s stated objection to same. I can imagine only one place the system might get energy to maintain temperature from at this point, and will have to read more to see if that’s what the authors mean and how they imagine this happens.
But my not understanding this aspect of their model doesn’t prove it it violates conservation of energy — rather, that tonight will begin further thought.
Willis, when I referred to your walkback it was because in your main post you definitely claimed there are 5 free parameters and now you are allowing that there may only be 4 (even though you think the correct number is still 5). So that is a walkback and I was not out of line for saying so. I certainly understand your broader position that 4 tunable parameters (2 pairs of linked) is still too small with such a small data set.
You are too quick to conclude that if someone takes issue with something you say, it necessarily means they have discarded; lock, stock, and barrel; everything you are saying, or that they are going to disappear off to your intellectual opponents’ blog to engage in unreasoned, closed-minded rah rah cheering sessions: in fact, it’s a little arrogant with a touch of petulance.
But I’ll overlook that because I can do that too.
My main purpose here is curiosity and wonder with a desire to figure out what’s going on, and also to support sound policies as a result.
But mostly the first thing. In that regard, I suspect you, I, Nikolev, Zeller, and of course out esteemed host, have something in common.
Christoph Dollis says:
“…I’ve certainly been thinking about the conversation of energy…”
I have conversations about energy all the time.☺
Ha ha.
Well spotted.
I think today I have had a major breakthrough in a sociological aspect of the N&Z phenomenon that has been puzzling me, which is why the followers of N&Z find their claim that their theory follows from the ideal gas law to be so compelling when it is so clearly a falsehood to those of us more familiar with using the ideal gas law.
I mean, their argument essentially consists of writing down PV = nRT, saying “and then a miracle occurs” and then writing down the four-parameter fit that they have used. How could this possibly be at all convincing?
Why such logic should prove so compelling to some has been mysterious to me until I realized that the answer was right under my nose: For the last 5 quarters in a row now, I have been teaching the quarter of introductory physics where we do a little bit of thermodynamics, including the ideal gas law. One particular exercise we do is present the ideal gas law and the law that E_th = (3/2)nRT where E_th is the thermal energy of the (monatomic) ideal gas, have the students predict what will happen under certain scenarios, and then have them run a kinetic simulation of an ideal gas to test their predictions.
There is one scenario that invariably trips up nearly all of the students: We tell them that an ideal gas has the number of moles n doubled while keeping the volume V and thermal energy E_th constant and ask them what happens to the pressure. They nearly all predict that the pressure doubles and are then surprised when the simulation shows them that the pressure doesn’t change.
I then go over this scenario with them in detail, first presenting what I assume to be their incorrect reasoning, namely that pV = nRT implies that p must double if n doubles and V remains constant. I explain to them that in fact they can’t draw any conclusion from the ideal gas law alone because they don’t know what happens to T. I then explain to them how they first have to look at the equation E_th = (3/2)nRT, from which they are forced to conclude that T gets cut in half (since this is the only way for E_th to remain constant when n doubles). Only then, once they know what T does can they use pV = nRT to uniquely determine what p does (i.e., that it remains unchanged).
The fascinating thing is that even given the simulation, the explanation of the error in their logic, and an explanation of the correct method of solving the problem, if I give the students either the exact same problem or a slight variation on this problem a few weeks later (like on the exam), the majority of them will STILL get it wrong!
The desire to just look at the equation pV = nRT and determine from it how the pressure changes when you change n is just too powerful to be resisted for most students and this is in a case where (unlike the case with N&Z’s theory) the students have no preference for what they want the answer to be!
Like any “successful ” piece of pseudoscience, N&Z’s theory relies on fooling people in a way in which they are very easily susceptible to being fooled.
Christoph Dollis says:
February 10, 2012 at 6:25 pm
This is altogether typical. You have accused me of “walking back” on my statements and claims, and said that I made a “stunning admission”.
I asked you for specifics, saying:
Your reply? You “haven’t even looked into what they’ve done” … I certainly hope you realize how pathetically stupid that sounds.
I also hope that next time you want to be involved in one of my threads, you do your homework first, so you don’t look so foolish. This is not your usual blog, we expect people to follow the bouncing ball and, you know, actually READ BEFORE COMMENTING.
w.
Christoph Dollis says:
February 10, 2012 at 8:41 pm
Oh, learn to read. I said they had five tunable parameters. I still think the same. TB or some other charming fellow said no, the fifth one was not tunable. I said maybe not, but it sure looked tunable to me. TB couldn’t explain where they got the number. I still think it was tuned like the others.
And that is neither a “walkback” nor a “stunning admission”.
w.
kuhnkat says:
February 11, 2012 at 10:11 am
kuhnkat, as the poet said, you seem to you have had the experience but you missed the meaning.
Joel didn’t say it was just a curve fitting exercise. N&Z said that.
You see the part about the “regression fit”?
It was verified by the fact that when I said it was a fitting exercise they didn’t say “no, we derived it using the following formula” or anything like that.
HAHAHAHAHAHAHA
w.
PS—Assuming that it is the custom where you come from to end your emails with wild, maniacal laughter, I have done the same to show solidarity …
Willis,
You’re just being an arrogant ass. That was apparent, but I wanted to give you the benefit of the doubt that you were a bit more than that in this situation.
There’s been a lot of writing back and forth on this and a lot to look into. No, I haven’t looked into every aspect of it at this point: I’m reading a blog and I made a comment on it. I am not, in fact, one of the scientists involved on either side of the debate, a fact which might have escaped you; this is not an international science conference. People with various backgrounds and interest levels have their say on a blog, where you have chosen to place your work.
I’ve commented off and on here for years, I read your post, and I didn’t read all of the then 490 comments before commenting, anymore than you read the very few new comments before replying as in the excerpted portion above … as evidenced by your next comment, which you begin….
Is this a big deal? Not really. It’s normal on a blog. But then I’m not whining about it either.
Let’s call a spade a spade here and say, based on simply the words you’ve written, your work looks sloppy.
In the comments you write this:
But in your posts, you write this:
This:
This:
This:
This:
And further, relate a story about Enrico Fermi criticizing an equation given to him with five tunable parameters, as you call them, being far many in that case.
Does our post give the distinct impression you are talking about five tunable parameters?
I’d say it does, and in light of that, when you acknowledge in the comments:
This is obviously comes across as a walkback, a retraction, from what you said in your main post. Regardless of whether you now maintain there are four or five and your arguments for preferring to think of it as a fifth parameter, I was commenting on what you yourself wrote in the excerpt immediately aboved as compared to what you wrote, definitely, your own post.
Accordingly, Ned and Karl were justified to write something along the lines of this:
… before you went on to demonstrate that x = x.
I’m sure that got your goat, and you continue to be emotive about it, but that’s hardly my fault.
However right you may ultimately be — and as I said, I am still thinking about this and reading further … as an interested layman following a scientific discussion and commenting on a blog where I’ve done so for years — you come across as someone who is covering his rear after mistakes were revealed. If that isn’t the impression you wish to convey, you may want to adjust how you communicate in response to criticism.
Or not. Carry on, sir. You’re being so convincing this way.
*Does your post give the ….
Christoph: Yes, Willis was wrong to think of that particular parameter as a free parameter (and I said is much in my first comment in this thread: http://wattsupwiththat.com/2012/01/23/the-mystery-of-equation-8/#comment-873837 ) But, does it make it that much better that they only have 4 free parameters?
What N&Z are trying to do by attacking any mistakes that Willis has made is distract you from the fact that their “theory” is utter garbage: It consists of a hypothesis that does not even satisfy conservation of energy together with a meaningless empirical fit to data and some nice words about the ideal gas law that have nothing to do with anything, along with various complete fabrications about how climate scientists have ignored the ideal gas law and thermodynamics when in fact the law is discussed in the beginning of nearly every textbook on climate science and when they are the ones who are doing violence to basic thermodynamic laws.
Oh yeah, and add to that: They claim that climate scientists have not considered convection correctly because of some vague silly complaints about the details of how it is put into climate models but then they themselves put convection into the simplest model of the radiative greenhouse effect in a way that drives the lapse rate to zero and then marvel about how this eliminates the radiative greenhouse effect when they would have known that if they had just read an elementary textbook in the field that they claim to be revolutionizing.
Christoph Dollis says:
February 11, 2012 at 2:09 pm
Naw, I’m just being accurate. It does sound stupid when you say you haven’t looked into what they did and yet you are foolish enough to make various flat statements about what they did. Sorry you don’t like it, bro’, but it sounds stupid as hell.
More to the point, you’re gonna seriously sit there and argue that the important issue was whether they had five or four free parameters? That’s what you think is important here, whether it’s five or four parameters?
Look, Christoph, you yourself claim ignorance as though it were a prize, saying you “haven’t even looked into what they’ve done” … and yet you want to be a nitpicking jerk about what they did? How does that work?
Come back when you have looked into what they’ve done, Christoph. I have no time to discuss this with people who proudly announce that they haven’t done their homework.
w.
PS—It’s not important, but N&Z refused to explain how they calculated the fifth number. I cannot calculate it from any combination of the data they provided. As a result, I will continue to believe that it is a free parameter until they or someone else shows it is not one.
About what you wrote, jackass. Where did I make these various flat statements about what they did, except when commenting on the discrepancy in your claims as posted and commented by yourself?
Never once did I do this. On the contrary, I took great pains to allow that you could be right overall. Show me once where I made this argument. Give it a shot. Because I can show you where I allowed for the distinct possibility of the opposite.
But this is the sort of if not dishonesty, emotionally biased “see what you expect to see” thinking I’m observing from you, like where to add unwarranted weight to your insult, you cut off my quote in an effort to take it out of context and change its meaning:
When I said:
It’s a subtle and important difference. Clearing I’m still gathering and processing knowledge, but your error was obvious, pointed out in the first comment of this thread, as you know, and I commented on it. Since then, you’ve even walked back your walkback.
[trimmed. Robt]
Willis, you are not so great at engaging a skeptic who is gathering information from different sides, including from yourself. The conflicting info written by yourself is what I commented on. Now I am considering Ned and Karl’s work generally and am thinking about it. I have concerns about it, such as all-important conservation of energy, as I stated above — also the possibility of errors in the math, rendering it less than meaningful.
However, I’m not going to conclude you’re wrong because you feel like playing the part of a noticeably defensive person, but certainly it doesn’t add anything to one’s confidence in your work.
Christoph, let me recap the bidding. You walked in the door and said
Christoph Dollis says:
February 10, 2012 at 3:30 am
You might not have tried to be a jerk … but yes you can, and you succeeded. My notation is entirely understandable. Sure, I could typeset it all in Latex … but since everyone but you can read it like it is, why should I?
So you started out by bitching and whining about my style. You immediately doubled down on your foolishness by saying:
Christoph Dollis says:
February 10, 2012 at 4:57 am
When some random dorkhead wanders into the discussion and starts out by accusing me of “walking back” with a “stunning admission!”, complete with gratuitous exclamation point, you can bet I’m going to slap him across the face. That’s no way to start out, by walking into a man’s house and doing your best to insult him.
Now you are wimpering about
You are not here to “gather information”. Your actions have shown clearly that you are here to insult and harass me over minor, petty bulldust. I have no intention to “engage” with anyone who acts like that.
w.
PS:
You happy now?
It is quicker and easier to visualize.
After 6 weeks of N&Z silliness, David M Hoffer, tallbloke, and the rest of the gang still don’t understand the basic question they are trying to answer! They keep trying to come up with sneaky ways to make the Earth warmer than 255 K ( http://tallbloke.wordpress.com/2012/02/09/nikolov-zeller-reply-eschenbach/#comment-17227 ) without understanding how the 255 K constraint comes about not because the sun doesn’t provide sufficient power to push the temperature higher if the surface of the Earth somehow were to not radiate energy back into space, but rather because of the fact that the Earth’s surface does radiate according to its temperature (and all such radiation escapes to space if the atmosphere is transparent to this radiation).
from a post on tb which does not seem to be answered:
Equation 2 seems to sum TSI and Cs then spread it round the globe. However Cs is already global so should not be so spread it must be constant over the surface. Inconsequential but physically wrong.
I do not see where the continuous downwelling radiation is handled in the equations. Like Cs this is day and night 200 watts so should not be spread equatorially although it does taper off polewards.
the 200Watts was measured here SGP Central Facility, Ponca City, OK 36° 36′ 18.0″ N, 97° 29′ 6.0″ W Altitude: 320 meters
http://www.arm.gov/sites/sgp
Eq 1 uses ap= Earth’s planetary albedo (≈0.3).
Eq 2 uses agb=Earth’s albedo without atmosphere (≈0.125),
Why the difference both assume atmosphereless planet?
I think that Ned Nikolov has, in his latest posts on Tallbloke’s, crossed the line into what I can only take to be deception:
Nikolov continues to repeat this falsehood far and wide even though nobody has challenged my contention that the reason why N&Z got rid of the radiative greenhouse effect by adding convection (in contradiction to the rest of the scientific community) is that they added it in totally incorrectly. We know that this is how they did it because they tell us they added it in incorrectly when they say, “Equation (4) dramatically alters the solution to Eq. (3) by collapsing the difference between Ts, Ta and Te and virtually erasing the GHE (Fig. 3).” I.e., they tell us that they added in convection in a way that leads to the completely unphysical result of an atmosphere isotropic with height, i.e., with zero lapse rate. And, any elementary climate science book would tell them that this would indeed eliminate the radiative greenhouse effect (e.g., p. 148 of Ray Pierrehumbert’s textbook).
When someone has been repeatedly told that they did something wrong and they refuse to even address this contention but just continue to make their false statements, I think it says something about them and whether or not they can be trusted at all.
Joel Shore and others repeatedly refuse to address any proof that he and the climate community are wrong. A planet’s surface temperature is dependent on the mass of the atmosphere and surface area and mass of the planet. CO2 and back radiation have nothing to do with with it.
Do they offer proof that their position is correct?
BigWaveDave: We have talked about the overwhelming evidence that our view is correct in these and other threads.
First of all, there is the empirical evidence: The Earth’s surface is emitting ~390 W/m^2. The Earth as seen by satellite from space is emitting only ~240 W/m^2 and the spectrum of that emission matches that of the radiative transfer models that compute the line-by-line absorption of radiation in the atmosphere, with the telltale absorption lines corresponding to water vapor, CO2, methane, etc. (There are also spectra taken of the sky from the Earth’s surface.) The entire field of remote sensing is based on the correctness of the physical equations used to calculate radiative transfer in the atmosphere!
Second of all, there is the fact that there is a law in physics called Conservation of Energy. There is overwhelming support for this law. Nobody has been able to explain using correct principles of physics how N&Z’s hypothesis could possibly satisfy Conservation of Energy…and they won’t be able to because it can’t satisfy it using any presently-known physical phenomena in our universe.
Does this constitute “proof” that we are correct? No, because there is no “proof” in science. It is inductive by nature. However, it does constitute absolutely overwhelming evidence. It is true that we cannot prove that N&Z are not correct…but we also can’t prove that the Earth isn’t only 6000 years old (or flat). We can’t prove that the next time that I drop an apple, it won’t fall up rather than down.
If you want to believe N&Z’s nonsense because of your ideology, then fine. Just don’t pretend you are basing that on scientific evidence because you are not.
…And, just to follow-up, we now know some of the major blunders that N&Z have made, in addition to being unable to explain how their ideas obey conservation of energy:
(1) They think it is a miracle that one can fit 8 pieces of data (some of the data points being rather trivial or having large uncertainties and questionable sources for the numbers) with 4 parameters.
(2) They don’t understand even the most basic facts about how convection operates in the atmosphere and have made a gargantuan error in putting it into a simple model and then marveled at the fact that their way of doing this gives the answer that any climate scientist…or even people outside the field who have read parts of basic climate science textbooks…could have told them they would get by making this mistake.
(3) In part 1 of their Reply to Comments on their theory, they show that they don’t understand how to apply conservation of energy to a system that is not isolated. Hence, they mistakenly believe that a rotating planet with a more uniform temperature distribution must have the same average temperature as a non-rotating planet with a very uneven temperature distribution. This is another incredible blunder that shows that they haven’t even a clue of what they are talking about.
Anybody who doesn’t run as furiously and fast away from N&Z as possible is basically telling the scientific community (and even the more scientifically-oriented members of the AGW skeptic community), “I am gullible and will believe absolutely anything, no matter how silly, that confirms what I desperately want to believe.” It looks like you have just added yourself to that list.
Over at tallbloke’s, Roger is continuing to allow Ned and Zeller to promulgate the blatant falsehood that adding convection gets rid of the radiative greenhouse effect. The following is something that he has refused to publish in response to his comments on what I said here: http://tallbloke.wordpress.com/2012/02/09/nikolov-zeller-reply-eschenbach/#comment-17513
Roger, the substance of your response is this statement: “The pressure and consequent thermal gradient is not included in equations 3 and 4 because it is not required for the purposes of demonstrating the inadequacy of radiative activity to account for the GHE (or the lapse rate), this is why the conceptual system under consideration is isotropic.” So, let’s consider this claim critically.
The claim that Nikolov and Zeller make is that the way in which convection is dealt with in the climate models is incorrect for some detailed numerical reason. They present no evidence to directly support their claim that the way in which the models handle convection leads to any unphysical results (such as an unphysical temperature distribution); rather, their sole piece of evidence is that when they add convection to a simple radiative model of the greenhouse effect in a way that they deem to be superior, they see the radiative greenhouse effect virtually disappear.
So, the question becomes: Is their method of handling convection really superior? As I have noted, their method of handling convection leads to a clearly unphysical distribution of temperature with altitude, i.e., it drives the lapse rate to zero by forcing the temperatures T_s and T_a to be essentially equal (as they themselves note). So, in contrast to the case of the way that convection is handled in the climate models, here we have unambiguous evidence that their way of handling convection leads to what is clearly an unphysical result. Hence, on the face of it, it seems rather unlikely that their way of handling convection is superior!
However, to get more evidence on this point, we need to address this question: Does the temperature distribution with altitude matter for the radiative greenhouse effect? The answer to this question can be obtained in two ways:
(1) We can simply look in the literature or in textbooks. For example, if we look at the textbook by Ray Pierrehumbert, we find a discussion that is summarized by the following sentence on p. 148: “The key insight to be taken from this discussion is that the greenhouse effect only works to the extent that the atmosphere is colder at the radiating level than it is at the ground.”
(2) We can study different ways of adding convection to the simple radiative model that they added convection to. For example, instead of adding convection so that tries to make the values of T_s and T_a to be equal, we can add it in such that it tries to make the atmosphere go to an adiabatic lapse rate (which, if we assume T_a is at a height of ~5 km would mean that T_s – T_a = 33 K). What one finds is that indeed now the greenhouse effect no longer disappears…but instead, that the temperature at the surface is ~33 K higher than the 255 K atmospheric shell, or 288 K.
So, what we have learned is that the temperature distribution with altitude is critical to the radiative greenhouse effect. In fact, it is both well-known and easy to show that forcing the lapse rate to go to zero (i.e., forcing the temperature at the effective radiating level to be equal to the surface temperature) gets rid of the greenhouse effect completely. Thus, what Nikolov & Zeller have shown is not new and their result is a direct consequence of the unphysical way in which they have put convection into the simple radiative model of the greenhouse effect.
As for Dr. Zeller’s response, what I have demonstrated is that it is the way that they have added convection to the model that is incorrect. I agree that if they add convection to a more complicated model in a way that is similarly incorrect (i.e., that allows the convection to drive the lapse rate to zero) then they will indeed get the same incorrect result that they got for the simpler model for the same reason. Guaranteed.