2011-01-13 08:52:19Pierrehumbert article on greenhouse effect in Physics Today
John Cook


Worthwhile article "Infrared radiation and planetary temperature" by Pierrehumbert  published in January edition of http://www.physicstoday.org/

Here's a link to the full article:


2011-01-13 09:28:17
Paul D


Thanks for the heads up and the link.

Looks good.

2011-01-13 09:57:41
Pete Murphy

I looked that this article  very helpful  but I haven't gone through all of the detail yet.

I'd draw attention to   two sections

(a)  P38  there are two  clear paragraphs  explaing the impact in simple terms and clearly avoiding  the common confusion over the Second Law of Thermodymanics 

 ...The planetary warming resulting from the greenhouse effect is consistent with the second law of thermosdynamic because the planet is not a closed system.  It exchanges heat with  a high temperature bath by absorbing radiation  from the photosphere of its star and with a cold bath by emitting IR into  the essentially zero-temperature resevoir of space. it therefore reaches equilibrium at a temperature intermediate between the two ....

The present Earth conditions  (etc) ......


(b)  I've always found the following statement  very hard to  wade through ...

P37  Energy balance and surface temperature

It therefore causes the surface temperature in balance with a given amount of absorbed solar radiation to be higher than would be the case if the atmosphere were transparent to IR .

2011-01-14 01:35:38
Paul D


I have read most of it and was thinking about figure 1.

It shows a layer model to show the radiative energy transfer through the atmosphere.

But it only shows emissions working their way up (the straight lines I assume is energy that isn't absorbed in the next layer up, but might be absorbed in the subsequent layer).

Surely you would also have a proportion of the 'transmission' downwards as well, that is the diagram should show straight lines pointing down. I assume that they are not included because the diagram would be confusing and the main interest is energy working it's way to space???

Maybe I have misunderstood the diagram??

2011-01-14 03:24:54

The radiation emitted downward by the upper layers of the atmosphere gets absorbed before reaching the surface. Only the energy emitted by the "first" layer reaches the surface.
Note that, contrary to the upward case where you have the almost perfect blackbody emission from the surface travelling through the atmosphere with discrete absorption bands, the radiation emitted by the atmosphere does not have the shape of the blackbody Plank function because it's weighted by emissivity. In other words, the wavelegth dependence of the radiation emitted by the atmosphere almost exactly matches the shape of the absorption coefficient of the nearby atmosphere; this is why it gets absorbed competely.
Not sure it's clear without showing some graphs.

2011-01-14 04:20:33
Paul D


Hi Riccardo,

Am I misunderstanding the straight lines in Perrehumberts figure 1??

Actually it might be clearer if someone explained what the difference is between transmission and emission in his diagram!??

He says "squiggly arrows indicate thermal emission, straight arrows indicate transmitted radiation".

I don't understand what the difference is. eg. surely an emission upwards from layer 1 is transmitted upwards to layer 2, they are one and the same thing??

I mean if you have a radio mast and transmit, you are emitting photons from the mast.

2011-01-14 05:15:24

Transmission is the part of light coming from somewhere else on one side of a layer and going through it to the next layer without being absorbed. Emission is the light emitted by a layer due to it's temperature and properties (emissivity); it's always there, independent on anything else.

Look at layer 1. Part of the IR coming from the surface gets absorbed and part goes through it to layer 2; the latter is the straight line. But layer 1 also emits IR, downward to the surface and upward to layer 2; this process is represented by the wiggles.
2011-01-14 09:14:18
Paul D


That's what I thought Riccardo, so my understanding is correct.

But that is where I don't understand why you wouldn't have some IR transmitted downwards (indicated by straight arrows) say from layer 3 through layer 2 to layer 1??
IR doesn't have a sense of the direction it is being emitted or travelling in.

Why would it get absorbed before reaching the surface, when if it were going in the upward direction some would pass through some layers?

Apologies for occupying your time!

2011-01-14 10:02:54

It's the difference of the characteristics of the emission between the surface and the atmosphere.
The surface emits like a blackbody with emissivity roughly equal to 1 at all wavelength and you have the full continuous emission. On the contrary, the atmosphere have discrete parts of the spectrum with high emissivity and others with emissivity equal to zero; the wavelegth dependence of the emitted radiation varies accordingly.

Consider first surface emission. It goes upward and it is absorbed where the absorption is high and transmitted where it is low. Overall you get that part of the emission can go through the atmosphere directly to space. This gives you the straight arrows. Try modtrans for clarity.

The emission from the atmosphere, instead, is zero where there is no absorption and high in the range of wavelength where absorption is high (emissivity and absorption go together, Kirchhoff's law). Hence the emission from the atmosphere is highly absorbed because it exactly matches the wavelegth dependence of the absorption coefficient. The result, to first order, is that emission from the atmosphere can go just from one layer to the next (no matter if upward and downward) while emission from the surface can in part go through.

The next step would be to consider the change of the absorption coefficient with temperature and density (altitude). When the radiation goes upward it finds progressively lower and narrower (in wavelength) absorption coefficients. In this way part of the (wider) radiation from below can go through upper layers as well. The opposite is true if radiation goes downward. You may have noticed in fig. 1 that the radiation from layer 1 in part goes through layer 2 toward layer 3 (straight arrow) while there's no transmission going from layer 2 through layer 1.

Please keep asking if you have any doubt; John set up these forums to allow discussions and to help each other better understand any climate related issue.
2011-01-14 10:24:30
Paul D


I understand absorbtion and emissivity go together.

But surely then in the diagram, emission upwards from layer 1 should be absorbed in layer 2, but Pierrehumbert has a straight line going through layer 2??

Oh actually I think it has just clicked!

You mean the composition (added: probably a poor choice of words!) of progressive layers upwards is slightly different (temperature, density etc) hence the absorbtion/emission characteristics changes and hence it is a one way path (upwards straight arrows only)??

Have I got that right?

Hope so!


2011-01-14 10:39:23
Paul D

I think I am understanding it now, but probably need to understand how temperature and density change emission/absorption properties.
2011-01-14 23:17:57

Crudely, the absorption peak widens on increasing temperature and/or pressure. If you do the experikment in the lab at constant absolute CO2 concentration, the widening is accompained by a lowering of the peak such that the total (integrated) absorption is constant. On the contrary, going up in the atmosphere you have lower pressure, lower temperature and less CO2 at the same time.
2011-01-15 07:32:29
Pete Murphy


Riccardo -  is all of the IR energy absorbed by each layer  emitted as IR radiation  ? Is any retained as 'thermal' ( molecular  kinetic  or rotational energy )  ?


2011-01-15 09:39:03

There are collisions by which a CO2 molecule gives away its extra energy. More, de-excitation by collision is more probable at high temperature and pressure and only relatively few excited molecules lose their energy radiatively.
Keep in mind that the reverse process applies as well, excitation of a CO2 molecule by collision. In steady state the two processes balance.

2011-01-15 20:42:59
Pete Murphy

Riccardo (or anyone else)   -   I'm thinking through this  and have managed to  confuse myself

    I fully understand  about the emission/absorption  setup  and how this is thought about with the  varionus layers.  

    What does it mean when we say the earths lower atmosphere has warmed . Are we saying there is increased IR radiation ( and no increase in  molecular kinetic energy and rotational energy) ?  I suspect I'm really asking how a mercury thermometer works  -  is it driven primarily by IR radiation or  by   'energy transferred by collisions'.


2011-01-16 04:37:09

The air in the troposphere, expecially in the lower part, is mainly warmed by convection initiated by the warmer surface. A smaller contribution is due to non-radiative de-excitation of the GHG molecules by collisions.
The IR radiation by itslef does not warm the air, its energy need to be converted to heat (kinetic energy). In some sense we may say that a molecule absorbs an IR photon and temporary stores its energy, waiting to "decide" what to do with it.
2011-01-16 05:12:42
Paul D


Re: PeteM. Scenario in sequence

1. Sunlight 'visible' photon hits ground
2. Ground emits IR photon upwards.
3. IR photon hits CO2 molecule
4. CO2 molecule absorbs IR photon and vibrates
5. CO2 molecule now has potential energy, but doesn't contribute to temperature
6. Time that CO2 retains IR photon energy is greater than the time that a oxygen or nitrogen molecule is likely to collide with it.
Hence it is more likely that the energy is passed on to another gas, than it is that CO2 emits the energy as an IR photon
7. CO2 molecule hits a Nitrogen molecule and the vibration of the CO2 molecule kicks the Nitrogen molecule, giving the nitrogen kinetic energy (increase in temperature) and the CO2 molecule relaxes.

With reference to point 6. Higher up in the atmosphere, the density will reduce, increasing the chances that the CO2 molecule emits an IR photon, rather than colliding with another gas molecule. Hence as you go higher, it is more likely that IR escapes, where as any IR emitted downwards in the higher layers is more likely to be absorbed (again).

The thing I have assumed though, is that at some point the Nitrogen molecule can transfer the kinetic energy it gained  to a CO2 molecule (or other GHG), causing the CO2 to vibrate and later emit (or collide and transfer energy again)???
(P.S I assume this is the case, otherwise how would the kinetic energy of the non-greenhouse gases be radiated to space??)

I realise this is simplified, maybe Riccardo can comment.

2011-01-18 11:02:59
Pete Murphy

The Ville  + Riccardo 

Just thinking this through  a bit further  and have two  last questions  . ( I'm  posing the first question in terms of Riccardo's point  "The air in the troposphere, expecially in the lower part, is mainly warmed by convection initiated by the warmer surface. A smaller contribution is due to non-radiative de-excitation of the GHG molecules by collisions.")

What does it mean  when we have heating through a warmer surface (since we're ruling out IR absorption/de-excitation as the main mechanism) .  Again, I'm trying to understand the part where  a molecule of CO2 or O2 or N2  close to the surface of the earth  gains more 'kinetic / rotaton energy'  by some collision with the Earth surface  (  I'm OK with the following convection)

How does the surface of the Earth  warm  by absorbing visible light  ( asume this is before a steady state has been reached) .    I'm assuming   somethig like  -  photon of visible light absorbed , photon of IR radiated ,  difference in energy is retained as  'kinetic energy'  of  Earth's surface (  =  rising temperature)  


If this is basic stuff I can read in  a text book  - please point me to the  right starting point . I did actually calculate some  absorption spectra many years ago  for some simple scenarios  but  never thought about it in  terms of  a complete heat exchange concept.

2011-01-18 20:31:46

At the microscopic level, air molecules in contact with a warm surface gain energy by collisions with the vibrating atoms of the surface. Collisions (almost) always dominate the exchange of energy at the atomic/molecular level, be it in solids, liquid or gasses (if not at low pressure).

Conceptually, there's no difference between absorption of visible and IR photons, it's always energy afterall. The difference is only in the microscopic process. Visible photons are absorbed by electronic processes, the photon is absorbed by an electron which will give the extra energy away (again) by collisions with atoms and/or radiatively. Vibrating atoms in the solid, in turn, emits IR photons so providing a mechanism to reach the radiative balance between the incoming visible energy flux and the outgoing IR flux. As far as the earth surface is concerned, evaporation and condensation play a role too.
2011-01-18 21:03:28
Paul D


I thought that IR from the surface warmed the first few metres of the atmosphere (it's thicker/denser at ground level), then convection was the main mechanism that moved the energy up from there. I may be wrong (often am!) but the point Riccardo was making earlier is that the convection mechanism for moving energy up through the atmosphere is more substantial than the radiative transfer mechanism.

The other main mechanism for transferring heat from the surface to the atmospheric gases is conduction, but that requires contact with the surface. I would think that the IR radiative mechanism has a greater capacity to initially remove energy from the surface, than conduction. Because it can reach a larger volume of atmosphere from the surface. However once the energy is in the atmosphere, convection is more efficient.

There is also the case of water evaporation (and other liquids/gases that end up in the atmosphere as a result of heating), which I guess is a mixture of methods.

 The following page I think confirms this, although it has 3 answers and the first one (Larry Krengel) appears to be doubtful!


Atmospheric Heat Gains
Short-wave radiation from the sun...............11.9%
Heat to atmosphere from condensation............14.4%
Heat to atmosphere from convection/conduction... 4.4%
Long-wave radiation from earth..................69.4%
Dale Bechtold, Meteorologist
Forecaster, National Weather Service
Weather Forecast Office, St. Louis, MO


BTW, a quick google of Larry Krengal suggests he is a aviation instructor!
Which sort of negates the credibility of the page title (ask a scientist).

2011-01-19 00:47:24
Paul D


 Found this lecture presentation which is useful. The last slide in particular:


Working out the percentages of heat transfers from the ground/earth.

I calculate the upward IR emitted by surface to be 76% (of the total energy transferred to atmosphere)

 Upward solar (shortwave) to be 5.5%
Latent heat 13.5%
Sensible heat 5%

The 76% for IR compares well with the 69.4% figure in my previous comment.

2011-01-19 00:58:28
Paul D


Actually what really annoys me.

Is that a lot of this stuff is really basic and a lot of people would like to understand it.
But why on earth does anyone have to search the web for hours and only find obscure sources of information that a school teacher or lecturer has put online??

If you go to some of the main sources of info, you find absolutely nothing!
You may find al sorts of stuff on conduction, convection and radiation, but none of it specifically addresses the earth/atmosphere and the stuff we are all interested in.

In fact during my searches, I found more skeptic sources than I did any other.

2011-01-19 02:39:44Climate Change 101
John Hartz
John Hartz


Perhaps SkS should create a new section, "Climate Change 101" to provide basic information about the key components of climate change. 

2011-01-19 03:44:48

We are risking to get confused because we're talking about different things.
I was talking about the mechanism by which heat is moved around in the surface-atmosphere system. The link provided by The Ville talks about single energy fluxes. The problem is, once more, the confusion between energy and heat. To call heat any kind of energy you need to have a temperature variation or a phase transition (change in internal energy); thus, the IR raidation by itself is not heat (yet). If a weird atmosphere always emits exactly the same amount it absorb, radiation has no direct effect on the temperature of the atmosphere (no heat).

Let's look at the radiation exchange between the atmosphere and the surface using Tremberth's graph. The atmosphere absorbs 356 W/m2 from the surface emission but 333 are sent back; we are left with 23 W/m2. In the mean time, the surface also "sent" 97 W/m2 by means of convection (which includes the latent heat release from moisture condensation). Overall, the heat trasnfered to the atmosphere by convection is about 4 times larger than ny radiation.

But this is not the whole story. Let's look at the overall radiative energy balance of the atmosphere. The radiative input is 356 W/m2 from the earth surface and 78 W/m2 from the sun; total=434 W/m2. The radiative output is 333 W/m2 toward the earth, 169+30=199 W/m2 toward space; the atmosphere looses 532 W/m2 radiatively and the balance is negative.
2011-01-19 04:04:03
Paul D


Hi Riccardo

I know IR radiation isn't heat (I learnt that in the last year and a half!).
I don't think we are getting confused.
Obviously there is also a downward longwave radiation.

But the point was to understand the mechanisms and how much each mechanism (in a simple scenario) results in an upward transfer at surface levels (ignoring downward transfers), assuming say the earth was warmer than the atmosphere immediately above.

2011-01-19 04:31:49Basic Definitions of Heat and Energy
John Hartz
John Hartz


As a non-scientist trying to digest the postngs on this comment thread, what are the basic definitions of "energy" and "heat"?


2011-01-19 04:58:31
Paul D


As far as heat and temperature in a gas goes, the molecules have to have kinetic energy.

That means they have to flying about, the faster they go, the higher the temperature and pressure.
If gas molecules are excited by photon absorption they are excited and may vibrate, that is potential energy. That isn't heat and isn't measured as temperature. However if they collide with another type of gas molecule they may pass that energy on as kinetic, hence causing the gas mixture to get hotter.

Actually Ben Miller did a good Horizon show (BBC TV) about two weeks ago about heat and temperature.

It is only in the last few months that I have learnt the difference between the potential energy of molecular vibrations and the kinetic energy of molecules flying about. I think most engineers would recognise vibration as a form of kinetic energy, at least that was what I did.

Another point is that the capacity of a molecule to absorb IR radiation (or any radiation) depends on the number of bonds, although the strength of the bonds and charges etc. Have an impact. More bonds between the atoms, the more likely the molecule will be a potent GHG. Oxygen and Nitrogen come in pairs and have a single bond (not IR sensitive). CO2 and Water vapour have two bonds. Methane has three bonds.

2011-01-19 05:00:51
Paul D


Actually Riccardo.

I think I have become confused about the difference between convection and conduction.



Added: That sounds terribly bad!

I mean I thought conduction was the contact of molecules and passing of energy between the two.
Where as convection was the tendency of warm gases/liquids at the macro level to rise, or rather the colder to sink.

Or can convection also occur between a solid and gas or liquid??}
eg. contact of gas/liquid molecules with solid.
I thought that was conduction!??!

2011-01-19 07:47:54

everyone understand what energy is, the problem is heat. We are taught that it is a form of energy but it is hardly detailed of what kind. You can think of heat as energy "transformed" into an increase of temperature or a phase change (solid-liquid, etc.). Following this concept, the IR radiation absorbed by a molecule is not heat until by some mechanism it is tranformed into kinetic energy (velocity).

The Ville
not true that the N2 and O2 do not absorb because they have a single bond or, better, they are bi-atomic; they do not absorb because they are composed by two identical atoms. If you take any other bi-atomic molecule with two different atoms, it will absorb IR at its specific frequency. It all stands on the dipole moment, something related to the presence of a positive and negative charge separation. IR radiation is an oscillating electric field and if there is no charge separation it "sees" the molecule as neutral and can do nothing on it. On the contrary, the electric field can move positive and negative charges closer together or farther apart, loosing energy (absorption) in the process.

You're right on convection/conduction but convection will not begin if we magically suppress conduction. Imagine a solid and a gas in contact at the same temperature; then you abruptly increase the temperature of the solid. How could the gas increase its temperature if not by conduction? Once the gas in contact with the surface begins to warm, convection starts.
2011-01-19 20:10:22
Paul D



I think my comment about molecule bonds was more a case of poor language. Although I did use the word 'capacity', the point I was trying to make was that the number of bonds is one of the factors that influence the potentency of a greenhouse gas. A complicated molecule has more ways it can vibrate, giving it a greater potential for storing energy.

Actually the next big question I have is the 'heating' of solids.
Does the same principle of temperature (one needs kinetic movement) apply to a lump of steel??

Given that the structure of a metal is a lattice/crystal structure, do the molecules/lattice have to have kinetic energy in order for the temperature to increase?
I suppose the fact that the structures are bigger in a metal, then molecular vibrations are going to cause kinetic movement because they will have an influence on the bigger structure due to the tighter nature of the structure??

eg. in the steel, the molecules are mostly in 'contact' (not the best word to use) with each other, hence any vibration caused by photon absorption will automatically cause kinetic movement.

Where as in a gas/liquid, the molecules aren't in contact, so they have to collide for molecule vibrations to convert to kinetic energy.

2011-01-19 22:20:49

In gases the simple picture of (average) kinetic energy proportional to temperature holds because molecules are free to move and there's no potential energy. In solids, atoms vibrate and you have a continuous "conversion" between kinetic and potential energy. In a microscopic picture, you can think in terms of total energy. Macroscopically (thermodynamically) it is called internal energy, which is a state function and depends only on temperature.

On passing, beware that temperature is a themodynamic quantity, it is strictly defined as an average in systems composed by a large number of particles. But we bad physicists use the same concept even for single molecules, while it really is a different way to indicate their energy. Astrophysicists, for example, often indicate the temperature of nebulae; it may be milions of degrees, but if you put a thermometer there it's deadly cold.
In this sense we may say that one molecule increases its temperature after absorption of an IR photon, but it's not really temperature in the usual sense. It's better to avoid this confusing concept when talking to non-physicists.
2011-01-20 04:35:16
Paul D


"In solids, atoms vibrate and you have a continuous "conversion" between kinetic and potential energy."

I think that is a better description of what I was trying to say!

2011-01-20 04:46:42
Paul D


I was thinking today, that the difference between potential energy and kinetic energy in the context we are discussing is a result of history. The science of measuring temperature was based on macro-scale science and observation with the human eye, the fact that for milliseconds some molecules might be holding energy and you couldn't measure it with a thermometer directly, meant it was un-noticed for a long time.
The fact that molecules have potential energy and are vibrating like mad, but a thermometer can't measure it (unless the molecule collides with the thermometer), doesn't mean it isn't important. It is energy in the 'system', which has to be accounted for.

Or am I going to be corrected again :-)

These days, such detail means the difference between a computer circuit working or not.

2011-01-20 07:03:07

Exactly, welcome to the realm of thermodynamics :)
2011-01-20 10:12:52
Pete Murphy

The reason I asked some questions about this  is because  I think there is great scope for confusion  about how IR  radiation actually leads  to  a change in  ' temperature  '  in the earths surface and  lower atmosphere.

IR radiation  is  transfer  of energy via E-M waves  (or photons if you're thinking Quantum Mechanics)  .

However if we have statements about   greenhouse gases keep the earth  33 oC ( Kelvin)  above what they would be  we've switched to thinking in terms of average kinetic energy per molecule  (  thinking of   gas temperature  being a representation of  average kinetic enery per molecule in gases -  solids/liquids have other constraints).

2011-01-20 19:25:05
Paul D

I think there is great scope for a series of articles, which for many would be basic physics, but for others would satisfy a great demand for knowledge.

I have to say that before 'climategate' I was happy to accept mainstream science at face value and basic analogies of the mechanisms, such as the classic, 'it acts like a blanket'.

So basically my knowledge (expressed in this thread and I guess expressed in my simulator project) has largely accumulated over the last year and a half or so. I probably understand more physics now than I did when I was a graduate!
2011-01-24 07:31:32
Paul D

I'm in the process of updating the description of the greenhouse effect on my blog as a result of this discussion and previous reading.
I have also updated some of the texts on my simulator.

2011-02-01 02:02:25
Paul D


I have updated some of my blog posts about carbon dioxide and 'climate science' as a result of these discussions:


For now I have actually removed the one about the greenhouse effect due to so many errors that were a bit misleading. I am planning to reinstate an updated version, however it needs a few diagrams I think and a major revision of the text.

One of the biggest problems is the level I'm trying to aim at. It has to be at a basic level.