Wednesday, February 06, 2008

A light dawns (and the sun sets)

G. Kramm over at #717 finally makes clear to Eli what Gerlich and Tscheuschner were trying to express in their Arxiv manuscript Falsifiation Of The Atmospheric CO2 Greenhouse Effects within The Frame Of Physics,. It remains wrong.

Kramm writes

“Please explain, for instance, in which way the tropopause region for which the so-called anthropogenic radiative forcing was estimated can warm the earth’s surface which has a temperature of about 65 K higher than the tropopause region. This is not possible without any kind of compensating changes. If it would be then we would be able to build a perpetual motion machine of second kind.”

Actually he wrote this a few times # 457

(4) Do you believe that the tropopause region for which the anthropogenic radiative forcing is calculated can warm the nearly 65 K warmer Earth’s surface without any compensating changes?

(5) Is it true that a perpetual motion machine of second kind can exist?

(6) Give the various citations of greenhouse effect explanations listed by Gerlich and Tscheuschner evidence that this greenhouse effect really exist?

(7) Is the current explanation of the greenhouse effect physically correct?

before the curtain fell from Eli's eyes. Eli also had the usual kindly personal missive from Tscheuschner. Needless to say the greenhouse effect does not require a perpetual motion machine of any kind, but as we will see the terminology used to describe the greenhouse effect can mislead the willing.

At equilibrium the Earth (including the atmosphere) must radiate the same amount of energy as it absorbs from the sun. If the amount of absorption is greater than the emission then something has to change so that the emission increases. As a general rule this means something has to warm up. Remember that radiation from the Earth in this context includes the atmosphere.

Without greenhouse gases the surface temperature would be about -20 C ON AVERAGE (right now the temperature in Barrow Alaska). Greenhouse gases absorb energy radiated from the surface, blocking a portion of the radiation that is emitted from the surface so it can’t escape to space directly. A further portion of the atmospherically absorbed radiation is later radiated to space, a portion is radiated down to the ground and absorbed there. GT&K object that this contradicts the second law of thermodynamics.

It does not. The net energy flow from the warmer surface flowing into the colder atmosphere is positive. There is no violation of the second law. We can see this using a detailed illustration of the energy balance

The surface loses energy to the atmosphere by convection (thermals) 24 W/m2, evaporation 78 W/m2 and surface radiation 350 W/m2. An additional 40 W/m2 radiates directly from the surface to space. 452 W/m2 is transferred to the atmosphere from the surface. 324 W/m2 is radiated from the atmosphere to the surface. On net 128 W/m2 moves from the warmer surface to the colder atmosphere. Net energy is moving from the warmer body to the colder body. The second law is fine with that. Energy is also moving from the hotter sun to the surface (168 W/m2) and from the hotter sun to the cooler atmosphere (67 W/m2). Again, the second law is fine with that.

So, how do greenhouse gases “warm” the surface. Note the scare quotes. One can formulate the process as the back radiation from the greenhouse gases warming the surface, but on net, radiation from the surface is larger than the back radiation. Greenhouse gases don't so much warm the surface as slow the NET energy flow from the surface to the atmosphere and from the surface and atmosphere to space. If you turned the sun off, the ground would cool because the NET energy flow is from the surface to the atmosphere and to space. But the sun rises and pours energy onto the ground. In that case the NET energy flow onto the ground is positive, composed of a solar and a back radiation component. This requires the surface to warm until radiative balance is achieved.

The point that GT&K miss is that what is warming the ground is solar radiation + the back radiation from the atmosphere.

Let us just look at the radiation part of the problem and ignore latent heat and convection (the numbers are not accurate but indicative. See the figure for more accurate numbers). If the intensity of the sun at the surface was 300 W/m2 without greenhouse gases the surface would warm until 300 W/m2 were radiated to space from the surface. With greenhouse gases (for arguments sake) 200 are radiated to space and 100 to the atmosphere, of which 50 are radiated back to the surface and 50 to space. The sunlight is still pouring in at 300 W/m2 so the energy flow to the surface is 350 (300+50) in and 300 out (200+100) of which 250 make it to space. The net energy input to the surface which includes both the solar and the back radiation is +50. This will warm the surface beyond what it was at without the greenhouse gases until enough emission is generated that the radiation to space equals the radiation from the sun. The effect is a combination of solar radiative heating and the fact that the greenhouse gases block a portion of the long wavelength radiation to space and return it to the surface.

If one increases the amount of greenhouse gas, the absorption surface longwave radiation will increase, the amount of back radiation will increase, less radiation will escape to space until the surface and the atmosphere warm enough to restore the balance.

UPDATE: Flavius Collium has a better way of putting this (see comments)
I'd summarize it that the greenhouse gases "hinder the cooling of earth's surface".

Same if you use a resistor to heat a less insulated object vs a more insulated one - the insulation in both cases will always be colder than the object but the better insulated one will reach a warmer steady state.
Eli also corrected some minor misspeaks (see comments)


Anonymous said...

I'd summarize it that the greenhouse gases "hinder the cooling of earth's surface".

Same if you use a resistor to heat a less insulated object vs a more insulated one - the insulation in both cases will always be colder than the object but the better insulated one will reach a warmer steady state.

One could actually demonstrate that in a youtube video.

Anonymous said...

In the first paragraph below the figure, you write:

"Energy is also moving from the hotter sun to the surface (168 W/m2) and from the hotter sun to the cooler surface (67 W/m2)."

I think by "cooler surface" you mean "cooler atmosphere".

Regards, J.

EliRabett said...

Thanks both for the comments, I've made the correction and moved Flavius' comment into the conclusion. It should (IEHO) become the default option

William M. Connolley said...

I think the easiest explanation is:

the GHE makes the sfc warmer because it receives radiation from both the sun and the sky.

There you go, thats it. The rest is just detail.

Timothy Chase said...

I have seen the second law argument made at Climate Audit, and more recently again at Real Climate by one of the Climate Audit regulars. The idea that the second law of thermodynamics has somehow just slipped the minds of the entire climate science community is for some reason quite appealing to them. No one has brought up a tornado in a junkyard as of yet though.

Unknown said...

I'm surprised anyone had the fortitude to plug through all 100-ish pages of that dribble. I noticed it was a pile of crap when in the first chapter the authors sta,rted throwing around the diffusion equation without any reason to do so (the greenhouse gas theory of climate has very little to do with the heat diffusion) and began accusing every person who has studied radiative physics with confusion reflection and absorbtion/emission.

Nothing says "we're trying to convince a lay audience we know what we are talking about" more than the inclusion of a whole bunch of random equations!

Oh well, for those of us who have taken classes in quantum mechanics, G&T was never going to convince. However, for lay-readers, Eli's take down is certainly a nice one. Not bad for a bunny!

Dano said...

I have seen the second law argument made at Climate Audit, and more recently again at Real Climate by one of the Climate Audit regulars. The idea that the second law of thermodynamics has somehow just slipped the minds of the entire climate science community is for some reason quite appealing to them.

Two recycled tactics in one - bonus!



Marion Delgado said...

very well put, flavius.

with the usual proviso that it's not like an actual greenhouse, per se, or precisely like a ceramic or rubber or paper insulator, but a more complex mechanism achieving, on balance, a somewhat similar effect.

Marion Delgado said...


"The idea that [FILL IN WELL-KNOWN SCIENTIFIC FACT, LAW OR PRINCIPLE HERE] has somehow just slipped the minds of the entire [FILL IN THE SCIENCE] community is for some reason quite appealing to them."

It is, isn't it.

Anonymous said...

Arguments simply don't get any dumber than the claim that "atmospheric greenhouse effects" are "fictitious".

If there were no greenhouse effect, the earth would be subject to the same kind of wild day/night temperature fluctuations as the moon, with no atmosphere (+120C to -170C).

Last I checked, outside temperatures were a little more moderate than that, but maybe where I live is anomalous.

Anonymous said...

Anonymous-- Why do you think if there were no greenhouse gases, that the earth's day/night temperature variations would be anywhere near the variations observed on the moon?

While you may be right, it seems rather unlikely as there are big differences:

a) The earth's days are 24 hours long. Lunar days are roughly 28 days long. So, there is much less time for the dark side of the earth to lose heat by radiation before it face the sun and regains heat.

b) Even if the earth's atmosphere contained zero greenhouse gases, any temperature gradients would still result in convection which would transfer quite a bit of heat from the light side to the dark side.

c) Even if water, which exist abundantly on earth but not the moon, had no greenhouse properties, it would still have latent heat. So evaporation, condensation, freezing and melting would moderate temperature gradients on earth somewhat compared to the moon

So, lack of greenhouse effects on the earth would result in a colder surface than we see; that's not controversial.

Since radiative loses would be greater, we probably would see greater day/night variations than we do see on earth.

But it seems rather unlikely the earth's day/night temperature variations would be as large as we see on the moon!

Greenhouse effects are real-- but the fact that the earth's diurnal temperature variations are smaller than the moon's can be explained without them.

Anonymous said...

The Moon's average temperature is around –25°C (range about +125 to –175°C, give or take a few).

Without an atmosphere, the Earth's average temperature would be about ...? Possibly around –29°C, or thereabouts (ish); range +82 to –140°C.

I know, I know, I know ... but the Earth has an atmosphere.

But I think that's what Anon 3:19 pm might have been getting at, Lucia (which doesn't negate anything you've said).

Cymraeg llygoden

Anonymous said...

Lucky for anon and the rest of us the earth has an atmosphere. Otherwise, last time he checked outside temperatures, he'd have had to wear a space suit!

The mistake of confusing the greenhouse effect with all other effects of the atmosphere seems to be fairly common on both pro-agw and anti-agw blogs. The spin is just different.

I've read it used as anon does to 'prove' the greenhouse effect. I've also read it used to 'prove' using the lapse rate in a radiative convective calculation is wrong. This then leads to the proof that the whole idea of a greenhouse effect is wrong.

I've read other novel ideas proving or disproving some tenet of AGW by explaining what would happen on earth were more like the moon.

If you get to pick your favorite error, you can 'prove' anything! :)

EliRabett said...

An interesting point that has occurred to Eli is that if there were no greenhouse gases, convection would be much higher because the temperature differences would be much higher and there would be no clouds (no water)

Anonymous said...

Hm, what if there were no liquid water, how much water vapor would there be if the whole planet stayed below freezing (not just on average but at every point)?

I suppose you'd have to get rid of all the geothermal and volcanic sources to reach zero water vapor in the atmosphere, even with the rest of the planet under entirely below-freezing conditions.

Anonymous said...

lucia said: "If you get to pick your favorite error, you can 'prove' anything! :)"


So, what is the standard error associated with your regression coefficient?

Anonymous said...

If temperature gradients were larger (as might occur with zero greenhouse gases) then natural convection will tend to be higher.

The rotation of the earth also induced some convection; and convection itself reduces gradients.

So, the even if our atmosphere were pure nitrogen, we'd have much smaller temperature gradients on earth than on the moon, by virtue of having an atmosphere.

(Also, with respect to the T^4 issue, the lower gradients results in smaller radiative losses than a planet with large gradients-- all things being equal. So, I guess convection is, in some sense, warming-- but not in the way we mean by the greenhouse effect.)

Sorry, I didn't return to that thread, so I didn't read your question.

I thought I'd answered once when you asked about statistical significance. I answered based on the f test, which is just as standard as a t test or using uncertainty intervals.

The slope I reported in that comment is significant based on the F test, a t test, or just informally eyeballing the error bars on the slope. The third, is of course, the least useful way to do the test.

But, since you want to know the standard error on the slope, slope of "m= 0.0016 C/year" it's 0.0002 C/year.

If you'd prefer 2 sigma error, That's 0.0004 C/year, which would tell us that the slope is between

0.0012 < m <0.0020 with something like 97% confidence.

I'm sorry this sounds fishy to you, but 1880-2007 gives us lots of annual average measurements from GISS and Hadley. If you feel the need, the data are publically available. Download them, and do the computation in Excel or your favorite statistics pakage. It's pretty easy.

My one caveat, which I've stated before is that I did not account for the inter-annual correlation in the residuals in the fit. If you do those, you'll widen the uncertaintly intervals, though only very slight, and that correlation is very small.

FWIW: Your handwaving argument estimating what the bias in instrumental measurements based on uncertainty in IPCC projections of warming is fundamentally unsound.

You are making an apples to frogs comparison!

The uncertainty I am describing is purely an 'instrument' issue. (I use the term instrument loosely here. Both groups are both trying to report the GMST over land and ocean. )

The actual trend in GMST could be highly positive, highly negative or flat, grossly uncertain or very certain. That would not affect instrumental uncertainty in the measurements. Instruments can still have uncertainty, bias and error regardless of the trends in what we wish to measure. If you take the time to actually do the computation using numbers, you'll see that your idea just does't pan out.

Anonymous said...

Now that I have your actual number for the standard error, I can compare it to what I get for the regression.

You may consider reality checks "handwaving", but suffice it to say that when I see claims that something is significant "to the 99.9999999997% level" (as you satted), I am a bit skeptical, to say the least.
As for statistical significance, it turns out, based on the Fdistribution, the trend I described is significant to the 99.9999999997% level.

I was not basing my comparison for the standard error to an IPCC "projection", but instead to a regression done on the average temp anomaly data for the past century (and past century and a half, resp.)

I have done the regression myself on the GISS temp data and came up with a standard error of 0.0006C/yr, which is 3 times the value you quoted for the trend for the difference between hadley and GISS.

My intuition says that when you take a the difference between two data sets, the uncertainty in the slope for the trend line should increase relative to the uncertainty for the trend line of either data set itself.

But, ass you say, the only way I can know that for sure is to do the regression for the difference.

Stay tuned. but based on what i have seen so far, I believe you are mistaken about the significance of your result.

It is rare indeed that a scientific result is significant to 99.9999999997%.

With all due respect, that is just so much nonsense.

Anonymous said...


I'm glad to see you are going to break down and do the regression rather than relying on your intuition about uncertainty in the difference in the errors.

My regressions use data from 1881-2007. I did them in January so the 2007 temperature data may be preliminary.

Linest gives this for the differences:
0.0016 -3.1580
0.0002 0.3969
0.3231 0.0843
59.6521 125
0.4242 0.8889

You will recognize the 0.0016 as the slope, and the 0.0002 as the uncertainty in the slope, and 0.32 as the correlation coefficient.

The F is 59.65, df is 125, the number of samples is 127 and if you use the FDIST function, you will get the probability these results occurred by chance is 3.16822E-12.

FWIW: The values for the GISS data over this period alone are:

0.0062 -11.9902
0.0004 0.7691
0.660860704 0.163412896
243.5801133 125
6.504508474 3.337971841

According to EXCEL, and contrary to your intuition, the uncertainty in slope for the GISS data, measured in C/year, is larger than the uncertainty in the slope for the differences.

(Thank heavens the uncertainty in the differences is smaller. The opposite would imply that either GISS or Hadley is providing us with some really truly noisy, bad, uncertain data! )

Meanwhile, those who would like to eyeball graphs of the GISS data, the Crut data and the difference to see what their intuition tells them about how well straight lines fit the GISS Met data you examined, and how well it fits the difference in Crut-GISS Met data may do so. To see, click my post of the data.

Anonymous said...

There are some things in life that money can't buy.

EliRabett said...

Ah yes, thanks to Lazar for bringing a fresh load of spittle into the burrow.

Kramm steps on another rake further below. The loons are indeed in full flight with the coming of Spring.

Anonymous said...


I had never realized bunnies take such an acute interest in wallpaper design. But wouldn't the little ones prefer something pretty, with flowers or tropical birds of foliage or such?

This ain't even pretty :-(

- Martin

Anonymous said...

Hey can you give me a hand debunking a skeptic please? She says that it's wrong to say "without greenhouse effect, the temperature would be 40 degrees colder". She says it should be "without an atmosphere, the temperature would be 40 degrees colder".

So I said "Well what about Venus and Mars huh? Venus is really hot and Mars is really cold, and Venus has a carbon dioxide atmosphere."

And she said that Venus is hot because the atmospheric pressure is 90 bar. For some reason she thinks that pressure causes temperature! Have you ever heard of such a thing!

And she says that Earth temperature is also caused by atmospheric pressure. In other words, the planet would be hotter if the atmospheric pressure was higher, or colder if it was lower.

So I said come off it. Everybody knows the greenhouse effect is real!

Then she asked me what would be the temperature of the earth if the whole atmosphere was nitrogen? And what would it be if it was 100% carbon dioxide? What would be the temperature if the atmosphere was the same pressure but different composition?

Well how should I know? So how do I debunk this please? Obviously the GHE must be real, or we wouldn't be spending trillions of dollars on fixing it up. Is there any way to work out what the temperature would be if the atmosphere was 100% carbon dioxide, compared to 100% nitrogen?

And is there any way to work out what the temperature would be if the atmospheric pressure was half or double what it is now, but with the same amount of carbon dioxide?