2011-05-31 21:43:22Basics of greenhouse spectra


John has invited me to post this article off my blog (presumably as a blog post since it's not really a rebuttal) on greenhouse spectra. Any thoughs or comments before I go ahead?

2011-06-01 03:15:32


Two thoughts:

- It would be useful to show on, or near, the graphs which of the peaks belong to which of the molecules

- I thought the reason that CH4 was so powerful was that there is so little of it in the atmosphere, but there can potentially be much more (methane ices, decomposition products, etc.). Since the CH4 GHG effect is logarithmic (as it is for CO2), a small increase in CH4 has more impact than would the same increase for CO2.

- Another point, hinted at but not fully explicated, is that the most important contribution to the GHE happens at the highest altitudes; and H20 quits at about 10 km, but CO2 goes on up to 100 km.

2011-06-01 08:25:53


"all molecules capable of absorbing a photon near a peak in the graph have absorbed a photon, meaning any more such photons can pass straight through."

I don't think it's correct, the molecules rapidly de-excite.

Also, the peak absorption coefficient in the CH4 band is similar to that of CO2 at 650 cm-1, a few m2/kg. In any case, they can not be directly compared in this units of measure.

2011-06-01 23:46:32


I noticed the point that Riccardo mentioned also, but then couldn't find it again.

It sounds like the old "CO2 is saturated" mistake.

2011-06-02 09:04:11


nealjking: each graph is for a specific molecule, so I don't understand your first point.

Riccardo: Over a short time-frame they emit but if you smooth out over long enough, effectively they are transparent to CO_2. If anyone can help with better wording here that would be useful, otherwise I'll try to think up something. I think you are missing the point about the differences in the peaks: that of CH4 is not near the peak of Planck emission for a body at the Earth's temperature, whereas that for CO2 is.

Thanks, all useful for clarifying the article.

2011-06-02 15:28:47
Chris Colose



I think this is a good contribution, and it seems like you may need help with some wording.  I've also put another image of radiant spectra (looking down from space) with the greenhouse overlap in this post you can consider adding on.

The efficiency of a particular gas at contributing to the greenhouse effect depends on its concentration and intrinsic absorption properties, its optical overlap with other gases, various "line broadening" (or continuum) mechanisms, and the vertical temperature profile.  The reason water vapor doesn't completely swamp the terrestrial greenhouse effect isn't really because if its short residence time (there's always water in our atmosphere) but because, as you note, there is still some leaky areas in the WV spectrum, especially near the 667 cm-1 (15 micron) vicinity, where CO2 is in fact a strong absorber.   The other reason is that the greenhouse effect is particularly sensitive to the presence of strong absorbers in the upper atmosphere, since the best way to reduce the Earth's energy loss is to preferentially emit more radiation from the coldest parts of the air column (this is why high clouds tend to have a warming effect).  Of course, for water vapor, there isn't much of it in the high atmosphere.  Water vapor could dominate the greenhouse effect almost completely in situations where the upper atmosphere becomes very moist, but this never happens under Earthlike conditions.

Concerning methane, as Neal mentioned, the reason it's thought of as being "a more powerful greenhouse gas" is really just because it "looks more powerful" at modern day concentrations (because it has a lower background concentration), and not anything to do with any intrinsic property of the gas.  Because, as you note, it absorbs far away from the peak of the Planck function at terrestrial temperatures, it is even less a good greenhouse gas than CO2.  It's just like adding 1 ppm of CO2 to a background of 200 ppm produces a larger radiative effect than 1 ppm to a background of 2000 ppm.  Adding 1 ppm of CH4 will produce a larger effect than adding 1 ppm of CO2.  I'm not sure how useful the popular quotes are about "methane being 20x a better greenhouse gas" or whatever people say, since the comparison is made on a molecule-for-molecule basis like that and the ratio depends strongly on the background of each gas.

Regarding the absorption coefficient graphs you have and the objections Riccardo raised about photons not being absorbed:  A good way to think about this is that since the vertical axis is the absorption coefficient k [in m2/kg], if there are M kilograms per square meter of absorber over a cross-sectional area, the light is attenuated by a factor of exp (-kM).  Thus, if you shoot a beam of light through a tube of air where the product kM is one, only ~37% of the original intensity will remain. Since k >> 1 in the opaque region near 15 microns, virtually no photons of this wavelength from the surface make it out to space, and it isn't until the stratosphere that the air becomes optically thin enough to let radiation out in this wavelength interval.  But an individual photon is constantly being absorbed and re-emitted (technically, it's a new photon, since photons are destroyed upon absorption), so it's really a cascading process of photons going up and down through the atmosphere being absorbed and emitted, and the time between each new collision (the mean free path) declines for wavelengths of higher opacity..  And, actually for tropospheric conditions, the molecules will collide with N2 or O2 on timescales much faster than it takes to de-excitate through direct emission, which allows us to define local thermodynamic equilibrium.  So, in the case of CO2, absorption is strongest in the line center and declines rapidly toward the wings of the absorption feature.  The wavelength domain over which the atmosphere is actually optically thick then depends on the concentration.  For sufficiently low concentrations, only the center is, but for higher concentrations, the "wings" start to become optically thick. 


Here is the graph I promised


2011-06-02 16:08:39


Chris, thanks for the comments. This shows the value of peer review. I thought I had explained some of this so I'll try to get it right. Do you mind if I lisf t some wording since you've taken the trouble to go to this level of detail? Happy to give acknowledgement.

How about you post the graph in a comment, then you can add in any detail on its provenance and significance?

Would you like a pointer added to your more detailed article?

I didn't go into some of the complications like pressure and continuum broadening because there's just so much you can say in one blog post. This is after all summary of part of a > 600 page text book. The upper atmosphere point is a good one too. In my next break from my day job (writing a paper about something completely different) I'll start making changes.


2011-06-03 00:20:09
Chris Colose


Sure, use what you like

2011-06-09 14:50:14
Chris Colose



To further enhance your post, I just made a plot up of the OLR as a function of concentration for CO2 and Ch4 (Tg = 280 K) in Python (taking advantage of the exponential sums radiation code originally developed by raypierre for his book).  It shows how CO2 is always a better greenhouse gas when comparing the same concentration side-by-side, since it always reduced the outgoing radiation more than the CH4 case, at a fixed T.

2011-06-09 15:45:51Fixed image
John Cook


Chris, your image was being tricky so I uploaded it to SkS.

2011-06-09 17:49:03not necessarily relevant to this
Otto Lehikoinen

as the absorption of specific wavelenght decreases logarithmically, if the CO2 concentration rises exponentially, one gets a straight line of T increase from CO2 alone, hence the additional effect of other greenhouse gases (H2O, CH4, well O3) rise the T more because they don't 'compete with CO2', put in additional feedbacks like albedo decrease and the exponential increase in T may become a possibility (Venus-syndrome), if the rise in CO2 stays at exponential levels (this includes natural sources like CH4 decomposition in the atmosphere). If the T rises to levels inhospitable to life that is a major (and can be pretty fast at the extinction events at least) sink to CO2, the dead planet follows (oxymoron). the geological weathering takes a long time, so larger amounts of carbon should be incorporated to life to fast reduce CO2 levels, but what's happening at the Amazon, and in Russia, Texas wildfires does not look like a positive development even in this respect.