RealClimate logo

Olympian efforts to control pollution

Filed under: — gavin @ 14 March 2009

There is a new paper in Science this week on changes to atmospheric visibility. In clear sky conditions (no clouds), this is related mainly to the amount of aerosols (particulate matter) in the air (but is slightly dependent on the amount of water vapour as well, which is corrected for in this study). The authors report that the clear-sky visibility has decreased almost everywhere (particularly in Asia) from 1973 to 2007, with the exception of Europe where visibility has increased (consistent with the ‘brightening trend’ reported recently). Trends in North American stations seem relatively flat.

There is another story that didn’t get as much press when it came out late last year but that is highly relevant to this issue – whether any of the efforts that the Chinese authorities to reduce air pollution ahead of the Olympics last year had any impact. To the extent that they did, they might point the way to reducing aerosols and other pollutants across Asia, but it might also reveal how hard it is to do so.

The press release and abstract for the Science paper link their results to the ‘global dimming’ trends we have reported on in the past, but it’s worth perhaps pointing out that previous studies (and the term ‘global dimming” itself) have referred to all-sky conditions. So that includes changes in clouds – which are obviously a big factor in how much sunlight gets to the surface. Looking at the clear sky conditions (i.e. only when there are no clouds) can help attribute changes to aerosols or atmospheric dynamics say, but since aerosols affect clouds (the ‘indirect effect’) as well as circulation too, it is only a partial estimate of the true impact of aerosols.

But getting back to the Olympics…. Monitoring of pollutants near the surface has improved enormously in recent years with the various satellite instruments now in orbit (MOPITT, GOME, OMI and TES for instance (sounds like a comedy revue team, no?)). These instruments detect specific frequencies where pollutants are known to absorb and so can give a birds eye view of where the pollutants are and how they are changing. Among other things, the satellites can detect ozone, NOx, SO2, the total amount of aerosols and carbon monoxide. Each of these have different atmospheric lifetimes and so can be used either to detect point sources (from pollutants that only last a short time) or long range transports of pollution (from the longer lived pollutants). NO2 (a big component of NOx – which lumps together NO and NO2all of the reactive nitrogen oxides), is very short-lived and so tells you a lot about local sources. Carbon monoxide has a longer lifetime (a couple of months) and so can show the long-range impacts. Many of these pollutants have related industrial sources (car exhausts, coal burning, industrial production etc) and so can be used as proxies for many other pollutants (such as specific aerosols) which can’t (yet) be directly measured.

What do the results show? The team at GSFC have released preliminary images from the NO2 analysis showing the before and during the pollution controls. In both images, Beijing shows up as a huge hotspot of pollution, but relatively, the levels during the Olympics were significantly smaller:

August 2008 levels were therefore about 50% less than a similar period the year before. Meanwhile values at other hotspots in China were steady or got even worse. So there was a significant effect, but the scale of the task was indeed Olympian.

342 Responses to “Olympian efforts to control pollution”

  1. 101

    #84 Richard Ordway

    Regarding the paper (Swanson, Tsonis Has the climate recently shifted?)

    I was able to find a draft, not the original so please forgive if there is a major point i have missed. But if this one is circulating and anyone wants to put it in perspective Gavin’s concise reply is wonderful. I thought I would try to add perspective to it as well for anyone that wants to post to blogs using that argument:

    1. The title of the paper is “Has the climate recently shifted?” It is presented as a question. It does not provide any answer of substance.

    2. The paper is based on the break in consistent warming 10/40 – 76/77 – 01/02. They are doing an analysis based on a perspective and considerate of resonant qualities (as appropriate in their context) as understood based on the scope of their analysis. However, the paper is not considerate of the relevant contexts such as

    a. Strong El Nino event 1998 set up the downtrend as
    b. we went into solar minimum in the Schwabe cycle, and were at the bottom of the PDO, and
    c. it is a non linear system and while climate inertias and momentums are intrinsic to the nature of the system, they shift based on imposition of the dynamics involved, such as,
    d. loss of summer ice loss in the Arctic which will allow more absorption of solar radiation, and
    e. oceanic thermal inertia, and
    f. return to solar maximum in the Schwabe cycle (the cycle averages 11.1 years but peak length is around 14 years, so we are expected to return to solar maximum within 5 to 6 years), and
    g. economic downturn may provide a lull on increased CO2 production, but it won’t remove the CO2 in the atmosphere already,

    3. They are assuming that state changes in the climate system are reliant on or attached to resonances, which is logical, but do they reasonably consider the magnitude of forcing components on the resonant qualities in relation to positive forcing of the system? Remember overall forcing is positive bias around 1.6 W/m2.

    4. It does seem to be a statistical analysis, and while they discuss significant breaks in temperatures trend, they don’t seem to account of the probable causes of those shifts such as 10’s to 40’s lots of coal burn but not as much aerosols in industrial output; 40’s to 70’s aerosols; 70’s to 2000 removal of some aerosols, lot’s more CO2. 1998 to 2008 El Nino peak to solar minimum (including minimum phase of PDO, etc.).

    5. The coupling theory is appropriate only to the extent that the coupling exists and one can as easily hypothesize that the increased storage of energy in the climate system may have resonances and decouplings at a faster rate than in an energy balanced system, thus rate of state change may be less likely to maintain it’s assumed inertia as they indicate in relation to time scale.

    6. They state “This cooling, which appears unprecedented… is suggestive of an internal shift of climate dynamical processes that as yet remain poorly understood.” While they do consider some factors known in climate they do not seem attentive enough in the paper to the short term resonant coupling factors with fairly well understood process of PDO, Schwabe cycle, energy balance short term inertia, oceanic thermal inertia, etc. I don’t know how to do the math on all of it but it seems they have limited the scope to short term and are presuming some sort of long term effect out of that.


    Again, it’s the weather vs. climate argument. Short term is weather, long term is climate. Their conclusions are merely suggestive and speculative, but to speculate that this is a state shift that “may persist for several decades”, though it would be nice, is unlikely based on the inconsideration of the aforementioned points regarding positive forcing bias and oscillations that are returning to positive phase.

    This is likely not an actual state change as suggested, but rather a resonant change based on the systemic cycles from certain negative phase components in natural variability. The amount of forcing in the system biases the state to warming. As the natural variables return to positive phase, it is reasonable to expect the state to return to it’s biased path, that of warming.

    System impacts tend to resonate based on magnitude of event, and resonant and coupling qualities such as transmissive capacity and resistance v. forcing and bias. These regulate magnitude of resonance and coupling capacity. Affects are coincident with capacity in these areas. i.e. the amount of change or shift has inertia, but is limited by contrary factors/forces.

    Baring another large volcanic eruption in between the tropics, we should reasonably expect a return to warming in the trend within 2 to 7 years as there is some negative inertia here. I am not a climatologist so there may be additional negative phase components I am unaware of. If anyone knows where I might find a list of all the relevant cycles I am referring to I would love to see it.

    I am suggesting that if we add the known positive and negative cycles and the resonant qualities/forces, as well as coupling to the parent climate system, we can reasonably project when we will return to the biased course of warming. If anyone can add to this I would like to learn more about other cycles I am unaware of, and what phase they are in.

  2. 102

    Re #93 & #100,

    It also sounds as if our unnamed friend hasn’t quite grasped the whole “water vapor” vs. “cloud” thing yet. Wrong thread, but it would be “good advice for young (climate) bloggers” to get that straight.

  3. 103
    William says:

    Does everybody know that the models you are discussing cannot replicate the actual temperature of the earth? They predict only anomalies from a calculated long term average. Regardless of how they are tuned and the parameters tweaked they cannot produce a result that reflect real world temperatures. Doesn’t that bother anyone?

    [Response: Hmmm… so if I find a model that has a mean SAT of almost exactly 14 deg C, you will henceforth trust in all of its projections? – gavin]

  4. 104
    MarkB says:

    Re: #101

    Richard Ordway,

    That looks like a good analysis, and some of those thoughts crossed my mind (2a, 2b, 4, for instance) when I read the paper. I want to run another point by this group to see if this critique is justified. The paper writes:

    “Assuming a mixed layer ocean depth of 200 m, an anomaly of roughly 1 Wm−2 should in principle have been sufficient to drive roughly a 0.25◦ C increase in global temperature since 2001/02”

    It seems like they using a higher-end estimate of GHG forcing in order to justify higher internal variability implied by observations. 0.25 C over 7 years translates into a warming rate of about 0.35 C per decade, higher than the average model projects.

    [Response: No one has claimed an imbalance of 1 W/m2 since 2001. The only estimate I am aware of was an average of 0.6 W/m2 over the decade 1993-2003 (supported by obs), and with a spot value of 0.85+/-0.15 W/m2 in the simulation for 2003. Given that year to year standard deviation in temperatures is around 0.1 to 0.2 deg C, and that you probably can’t neglect some amount of deeper ocean warming, these kinds of short term calculations are in the noise. – gavin]

  5. 105
    Ike Solem says:

    Re Vermeer:

    Reposting Steve Mcintyre themes on ” the 15 um CO2 absorption band” is a bit repetitive – and has already been discussed multiple times:

    CO2 absortion (1)

    CO2 absorption physics

    Water vapor and CO2 absorption

    In particular, you want to look at the figures:

    Sea level absorption spectra for greenhouse gases:

    Idealized blackbody vs. 11 km CO2 vs. sea level CO2:

    Oxygen absorption spectrum vs. distance above sea level:

    Now, compare those figures to the one that Steve McIntyre’s 15 um notion revolves around (linked on Climate Audit):

    Hmmm…. what does that say?

    Re#82, “Maybe I made an error (please point it out if so), but I’m surprised the water vapor positive feedback is so small.”

    To be polite, I think you could spend more time reading the literature on the water vapor feedback… but, for your convenience, here are some pointers:

    Yes, Richard Lindzen used to say that water vapor had no effect, and that the water vapor response in models was too high, but the fact is that water vapor feedback is coming out much like the models predict, with important variations in time and space – like the expansion of the subtropical dry zones.

    Or, look at this:

    The Radiative Signature of Upper Tropospheric Moistening, Science 2005, Brian J. Soden,1* Darren L. Jackson,2 V. Ramaswamy,3 M. D. Schwarzkopf,3 Xianglei Huang4

    Climate models predict that the concentration of water vapor in the upper troposphere could double by the end of the century as a result of increases in greenhouse gases. Such moistening plays a key role in amplifying the rate at which the climate warms in response to anthropogenic activities, but has been difficult to detect because of deficiencies in conventional observing systems. We use satellite measurements to highlight a distinct radiative signature of upper tropospheric moistening over the period 1982 to 2004. The observed moistening is accurately captured by climate model simulations and lends further credence to model projections of future global warming.

    By the way, wasn’t this post actually about reducing aerosol pollution? Can we perhaps get back to considering questions along the lines of what strategies are most effective at reducing aerosols? How about solar-powered vehicles, for one? For India, the solar-powered rickshaw is promising.

    Oct 2008 – The most common sight in India, the wooden rickshaw, pulled by people for a fee, could soon be replaced by modern, electric counterparts, as part of India’s effort to reduce the heavy pollution it is facing, as well as its dependency on fossil fuels. New Delhi has hosted this month the exhibit of the first solar-powered rickshaw prototypes, whose job is to relieve the clogged streets that characterize today’s India.

  6. 106
    Chris Colose says:


    on your reply concerning the Tsonis paper, I think you’re a bit too generous in your description if it. I’m not sure the authors understand what they are talking about. I don’t think its main focus is “about how the climate reacts to forcings” but rather is taking after the concept of “internal radiative forcings” although they don’t use that terminology. After a brief read, it doesn’t sound serious to me.

    [Response: I’ll concede that the paper in discussion seems a little hand-wavy, but it is a serious approach. I’m not personally convinced of its merit, but I’m willing to see where this goes. It’s still apparent that the reaction in the denial-osphere is completely out-to-lunch with respect to the merits (or otherwise) of this paper. – gavin]

  7. 107

    ziff — I get 20.5% from Essenhigh’s (2001) carbon dioxide scheme, and Essenhigh, let me remind you, is a denier, though a more sophisticated one than most. The fact is that CO2 accounts for 7 K of the Earth’s 33 K greenhouse temperature increment. It should also be noted that water vapor bands cover areas of the spectrum that CO2 doesn’t, and the more warming from carbon dioxide, the more water vapor in the air.

  8. 108

    Adam Gallon writes:

    A straight plot of “temperature” against “CO2 levels” isn’t a straight line, as I’m sure you’re aware.
    The relationship is poor, hence the need to add “forcings” by other atmospheric components.

    Actually, it’s a spectacularly close relationship. Remove the hyphen and post this into your browser:

  9. 109


    I didn’t criticize Dawkins because he summarized the work of others. I criticized him because I think his theory of gene selection is completely wrong. Ditto his penchant for sociobiology. I was a Dawkins critic long before it was fashionable. [edit]

  10. 110
    Martin Vermeer says:

    Ike Solem #104: I have no idea who you are arguing with — or what about. I have nothing in common with Steve McIntyre — trying to insult me, are you?

  11. 111
    Rod B says:

    re ziff’s reference in (87): the article says Mann’s paper was published in Nature (1998) without peer review. That sounds odd. What’s the situation? I’m not resurrecting the Mann “hockey stick debate — just curious about Nature publication rules.

    [Response: Not true in the slightest respect. – gavin]

  12. 112
    Mark says:

    BPL, [edit]

    It’s probably better to leave Dawkins to one side.

    [edit – these discussions are OT]

  13. 113
    Rod B says:

    Hank (99), my thinking was not near as philosophical as your defense. When ziff says his source says CO2 absorbs in only three bands and Mara says they’re full of it, that’s hyperbole, plain and simple. I won’t go as far as to say WRONG because Mara kinda clarifies it in the fine print.

  14. 114

    Martin Vermeer wrote in 109:

    Ike Solem #104: I have no idea who you are arguing with — or what about. I have nothing in common with Steve McIntyre — trying to insult me, are you?

    No, he didn’t intend that. He was responding to your post 94 (where you responded to Ike’s 78 and to Steve Reynolds’ comment 82) and he was also responding to Steve Reynolds’ comment — the same comment that you were partly responding to.

    In any case, I can see how it can get a little confusing when one a given comment is used to respond to different people, particularly when those people hold different views which one is responding to — even when you indicate in the text where you are switching from responding to one person to responding to the other.

  15. 115
    Steve Reynolds says:

    Ike#104: “I think you could spend more time reading the literature on the water vapor feedback… but, for your convenience, here are some pointers:…”

    That long list may have the answer to my relatively simple question somewhere, but I did not find all of it. This was somewhat useful:
    “They found that using the observed volcanic aerosols as forcing the model produced very similar cooling to that observed. Moreover, the water vapour in the total column and in the upper troposphere decreased in line with satellite observations, and helped to increase the cooling by about 60% – in line with projections for increasing greenhouse gases.”

    I think the 60% agrees reasonably with david’s model for water vapor increase when it says CO2 doubling direct sensitivity is 0.85 degrees for constant water vapor, and 1.26 degrees for constant relative humidity. What I did not find is why david’s model prediction is only 0.85 degrees for constant water vapor. Is that correct or not and why?

  16. 116
    veritas36 says:

    Clear skies – no clouds ???
    When does this happen now? As a sky-watcher (for pleasure) I’ve noticed that the bright blue clear skies are gone in the last 2-3 years. At best the sky is whitish blue, not ‘cerulean’. Jet contrails persist, criss-crossing each other in the sky as they widen long after the jet has passed by. Contrails used to vanish behind a silver jet almost as fast as they were created on clear days.
    I have questions: is anybody measuring a significant change in albedo? What is causing it (judging from photos on the net it’s worldwide.) If haze has increased, will it increase or decrease reflection of sunlight with the resultant change in global temperature.?

  17. 117
    Richard Ordway says:

    #106 Gavin wrote: “It’s still apparent that the reaction in the denial-osphere is completely out-to-lunch with respect to the merits (or otherwise) of this paper.”

    Some blogs do seem to be getting excited (inaccurately, but a fact nonetheless) over the Tsonis study. An example:

    “The latest peer-reviewed study in Geophysical Research Letters is being touted as a development that “could turn the climate change world upside down.”

    The study finds that the “Earth is undergoing natural climate shift.” The March 15, 2009 article in details the research of Dr. Anastasios Tsonis of the University of Wisconsin-Milwaukee.

    “We realized a lot of changes in the past century from warmer to cooler and then back to warmer were all natural,” Tsonis said. “I don’t think we can say much about what the humans are doing,” he added.”

  18. 118
    Hank Roberts says:

    > clear skies

    Perhaps someone at NASA reading this will recall some discussion of this — I recall reading sometime around the Spacelab/Mir years that the astronauts who went to the Moon had a very clear view of Earth from space, but since those years the atmosphere had become increasingly murky and no astronaut in then recent years had had the same crystal-clear view seen by the earlier space travelers from Yuri Gagarin up to sometime in the 1970s.

    The meteorologists also had a name for it, around that time I think it was being called the “midwest pall” — the general haze from the increase in coal burning at the time, before the Clean Air Act at least.

    Can’t find it with my usual 15-second research program online. I’m sure the optical astronomy people must keep track of this sort of thing too.

  19. 119
    Hank Roberts says:

    Rod, seriously, you’re making this up and not even realizing you’re fooling yourself by misremembering.

    Three specific _wavelengths_ were specified.

    Wavelengths. By number.

    A percentage was calculated from those three numbers.

    Not three bands.

    Catch yourself, man, nobody else can!

  20. 120

    #104 MarkB

    I’m not Richard Ordway ;)

    #104, #106 Gavin, thanks for the clarification.

    My opinion is it’s a semi-decent paper, but too limited in scope of consideration (hoping that is not an intentional bias) and without any reasonable merit on that point alone. Context too limited to justify conclusions.

    It seems to have some merit at a smaller scale than it seems to infer. Coupling, state changes, resonance are all well and good, and have a degree of relevance, but in the case they present, I seriously doubt enough to support their conclusions (suggestions/speculations), id est, “may persist for several decades”. Two 30 year shifts and one 10 year shift with no ties to the subsystem forcings does not cut it because they left out very important parts.

    Please pardon a bad pun, but there is an old saying that: Bikinis are great, they reveal a lot information, but still hide critical parts.

    I think it may be a great start though and hope they plug in the numbers of the critical sub-systems forcing, inertia, cycles, and weigh that with parent system forcing bias. Then maybe we get better at short-term climate inertia prediction?

    I’m hoping for a third paper that is more holistic and considerate of the major/minor systems and that they reach beyond the limitations of the current paper (and maybe add a few more names in there that understand those forcings). That could actually be useful.

  21. 121
    David B. Benson says:

    veritas36 (116) — I suggest the effect is due to ABC, Atmospheric Brown Cloud.

  22. 122
    SteveF says:

    Carl Zimmer has an interesting discussion of the Swanson and Tsonis paper, including comments by the authors on the reaction in the denialist blogosphere to their work:

  23. 123
    Maya says:

    hy⋅per⋅bo⋅le /haɪˈpɜrbəli/
    –noun Rhetoric.
    1. obvious and intentional exaggeration.
    2. an extravagant statement or figure of speech not intended to be taken literally, as “to wait an eternity.”

    So … if they were indeed trying to claim only 3 discreet frequencies, rather than 3 bands of frequencies centering on the ones cited (which as I said, was not clear to me from the quote), then they were indeed full of it, absolutely no hyperbole intended. :D If they were talking about bands centering on those frequencies, provided the claim was accurate, then it was merely sloppy rhetoric. Again, devoid of hyperbole.

    For the record, I’m fairly careful in what I say, particularly online. I may exaggerate somewhat for effect, but I am not given to true extravagances in my speech, either spoken or written, unless I am being obviously silly for the humor of it. If you wish to call someone out for such extravagances, you may wish to choose an easier target.

  24. 124
    Maya says:

    Richard, that was a good essay – recalling my own (ongoing) journey of understanding, I think they nailed it pretty well.

  25. 125
    Mark says:

    “Can’t find it with my usual 15-second research program online. I’m sure the optical astronomy people must keep track of this sort of thing too.

    All amateur astronomers are *really* pissed off with the last two years (at least). Lots of cloud, lots of bad visibility. Especially us urban astronomers. When there’s no cloud, the crud that now seems to be up there in its stead glows orange.

    Bloody annoying.

  26. 126
    Carrick says:

    Gavin, you may or may not remember that about a year ago I brought up the possibility that the recent plateau in cooling might be associated with recent increases in anthropogenic aerosol particles.

    Does this article tend to foster that notion, or is the amount of decrease in surface heating insufficient to explain the observed flattening of the global warming trend?

    [Response: Well, emissions inventories still haven’t been updated from 2000, and so we are still a little in the dark (so to speak). The big problem is that the global mean impact is a complicated balance between the decreases in Europe and the US, and increases in Asia. This data might tilt the balance towards Asia (and thus for a bigger cooling impact), but it isn’t definitive. Your idea remains within the realm of possibilities though… – gavin]

  27. 127
    Maya says:

    ::hands Hank a catcher’s mitt::

  28. 128
    Jamie says:

    Maya Hank, Rob,

    “The spectrum of heat absorption by Earth’s atmosphere contains hundreds of thousands of absorption “lines”. For carbon dioxide alone there are over sixty thousand lines.”

    Isn’t the resolution of ‘bands’ into ‘lines’ completely dependent on the scale at which one measures? Is there a single standard scale at which all electromagnetic spectra are universally measured? Does wavelength vary continuously, or in quantum steps? It seems to me that how many lines one finds must relate to the resolving power of the tool one uses. Is this wrong?

  29. 129
    Hank Roberts says:

    Jamie, it’s a description of what’s observed, and looking at the papers it takes more than a few sentences just to describe how they look. Did you read the papers linked above? Put those into Google Scholar and look for papers that cited them or for more recent work on the same subject. The search tool at the top will find several earlier topics about that question.

  30. 130
    Rod B says:

    Hank (119), the direct quote from ziff’s post is, “…carbon dioxide absorbs infrared radiation (IR) in only three narrow bands of frequencies…” (emphasis mine). It’s common usage to refer to a band by its main but general wavelength, e.g. 15um (but not 15.2378um.)

  31. 131
    Patrick 027 says:

    Re 80 – (ziff house)

    For substances/phases that undergo relatively quick physical and/or chemical reactions (H2O vapor, clouds, aerosols, … O3), atmospheric mixing cannot everywhere keep up with sources and sinks, and concentrations can be spatially variable, sometimes highly so. Regarding vertical variations, the general cooling up to the tropopause limits the water vapor concentration that can reach the stratosphere; variations in temperature, pressure, and UV will obviously be important to photochemical reactions and thus to O3.

    CO2 has sources and sinks at the surface allowing significant exchange between the atmospheric CO2 and other C reservoirs over a few years (but the “residence time” of any CO2 molecule should not be confused with the longevity of a change in CO2 amount in the atmosphere, which is considerably longer (there is a difference between the photosynthesis rate and the rate of net biomass accumulation, for instance). There is significant diurnal and spatial CO2 variability in the immediate vicinity of the surface in at least some location (under forest canopies, in cities, I think – don’t know the details) but CO2 in the bulk of the troposphere is well-mixed, with some seasonal variations, largest in the northern high latitudes (where much seasonal vegetation is found) that are small compared to the overall change in the last century –

    ( – The annual range of several stations is highest for Barrow Alaska (often close to 20 ppm), north of 60 deg N; next highest are Alert, Canada (farther north) and La Jolla Pier, California, between around 10 to 15 ppm; Baja California Sur, Mexico, and Cape Kumukahi and Mauna Loa, Hawaii, in northern low latitudes, are about between 5 and 10 ppm; near the equator, Christmas Island is around 5 or less ppm; the South Pole station has roughly 2 ppm, give or take, and the three stations in the middle-to-low southern latitudes (Cape Matatula in American Samoa, Kermadec Islands, and Baring Head in New Zealand)have little to no annual cycle.) –

    I would expect that variations in CO2 above the stratosphere and above vary less by latitude and have seasonal variations reflecting the tropical seasonal cycle, with reduced amplitudes; the relatively slow vertical overturning above the tropopause is driven by kinetic energy supplied from below via vertically propagating fluid mechanical waves (Rossby waves, gravity waves) – in the stratosphere, this is the Brewer Dobson circulation (I’m not sure if that name also applies to the mesospheric part), and involves rising in lower latitudes, drift to winter high latitudes and sinking in the winter polar region (some portion, at least, occurs in fits called sudden stratospheric warmings); the mesospheric portion involves rising in the summer high latitudes, drift to the other hemisphere, and sinking in the winter high latitudes.

    Obviously this motion is slow enough to keep volcanic aerosols (particularly from low-latitude eruptions) in the stratosphere for over a year, but for an increase in CO2 that has been going on for decades, the stratospheric and mesospheric concentrations will closely follow the troposphere. Generally, gases that are generally unreactive within the air, without significant sources in the air (not much CO; by molecules, much less CH4 than CO2), such as CO2, N2, O2 (much more abundant than ozone), are well-mixed not just in the bulk of the troposphere but in most of the atmosphere below the ‘turbopause’ (somewhere around 100 km up, in the thermosphere) – the fraction of atmospheric mass, and atmospheric optical thickness at most wavelengths, above the turbopause is very very very very small and (along with the mesosphere) can be set aside for calculating the radiative energy fluxes of the troposphere and stratosphere.

    (Above the turbopause, molecular diffusion dominates over eddy diffusion (macroscopic overturning and mixing), and the concentration of heavier molecules decreases with height relative to ligher molecules; at some point, atomic O dominates).

    (I don’t know how or if the concentration of CH4 changes with height within the stratosphere – I think it can be broken up by UV and whereever it is, it does eventually oxidize to CO2 and H2O over a decade or two, but whether the reactions are fast enough above the tropopause to affect the concentration significantly, I don’t know).


    Re 90 (Rod B) – The individual lines are broadenned enough so that, while the absorptivity may fluctuate by an order of magnitude (? – or more – depending on vertical position (pressure, temperature), etc – see book by Ray Pierrehumbert (further reference to come) ) over individual lines, the absorptivity at a line center (not too close to the strongest part of the band of lines) will be less than the relative minimum absorptivity between a pair of lines at some sufficient wavelength interval away – or at least this is the case for the band centered near 15 microns – the most important by far for CO2 in LW (wavelengths dominated by surface and atmospheric thermal emission) and more important than bands/lines in SW (solar-dominated) wavelengths.

    So I would think it could be called a continuum, at least for tropospheric and maybe lower stratospheric (?) conditions – but maybe there is a stricter definition of continuum that does not apply to this case.


    Re 83 (Maya) – see my Re 90 above; – a gasseous substance has individual absorption lines – for non-interacting molecules at rest, (or perhaps a single molecule averaged over time for many opportunities of photon absorption/emission), the line would be infinitely thin, thus having infinite absorptivity per unit wavelength within the line (though finite absorption cross section over a finite wavelength interval, proportional to the strength of the line) – except for ‘natural broadenning’ which as I understand it is due to the Heisenberg Uncertainty principle. The dominant broadenning mechanisms, however, are temperature dependent doppler broadenning (lines are blue shifted/red shifted by individual molecular motions, which are random, so the bulk effect is to spread the blur the absorption spectrum of the bulk material) and collisional or pressure broadenning, caused by molecular collisions. As I understand it, line broadenning takes the absorption cross section of a line and spreads it out over a range of wavelengths. Molecular interactions can also produce additional absorption lines, which is probably why the gap in water vapor absorption between about 8 and 18 microns fills in at sufficiently high specific humidity (PS this does not nullify additional radiative forcing from CO2, because it only happens at lower levels, so CO2 above such humid airmasses can still block some radiation from reaching space – see next paragraph). Line strengths can also be temperature dependent.

    (At local thermodynamic equilibrium, at any given wavelength and in any given direction (absorption from and emission toward) absorption cross section = emission cross section; absorptivity = emissivity. A cross section is the effective area, normal to (facing) a particular direction, with which an amount of substance can absorb or emit as a perfect blackbody. Emissivity is the intensity of emitted radiation as a fraction of blackbody radiation. Blackbody radiation intensity is only a function of wavelength, temperature, and index of refraction (the last not being important in the atmosphere or space). The reason the greenhouse effect works is that at any given wavelength wherein the air has some finite opacity, the absorption/emission cross section of air at some level hides the radiation (absorbing it) going in some direction coming from behind it and replaces it with it’s own radiation continuing in the same direction – which will be of greater or lesser intensity than that which it is hiding if it has higher or lower effective temperature than the source of the radiation it is hiding (the effective temperature being a weighted average over a path length of sufficient optical thickness that the emissivity approaches 1 – this includes the cold of space (can be considered a blackbody near absolute zero for these purposes) and the surface – radiation coming up from the surface can include some radiation emitted by the atmosphere that has been reflected by the surface because the surface is not a perfect blackbody, though it is not too far from being so in LW wavelengths).


    Re 78 (Ike Solem) – regarding stratospheric warming/cooling – Yes, the stratosphere cools in response to increased well-mixed greenhouse gases (CO2) – this is because the stratosphere recieves less radiation from below, but at the same time, becomes more ‘visible’ – radiating more energy upward and downward for a given temperature distribution. The cooling decreases that radiation, bringing the energy budget back into balance (reducing the tropopause level radiative forcing somewhat – often tropopause level radiative forcing is given for such an equilibrated stratosphere (but before tropospheric and surface responses) PS for the benifit of those who don’t know, the surface temperature tends to respond to tropopause level forcing because radiative equilibrium would be convectively unstable; convection tends to keep the troposphere vertical temperature profile near neutral stability to moist convection (where convection occurs – not in polar regions so much); the tropospheric and surface temperatures thus tend to rise and fall together in response to radiative forcing. There are some deviations – in particular, warming is concentrated near the surface at higher latitudes in winter because the surface is where the ice-albedo feedback occurs and the temperature profile is generally stable to convection (air gets heat advected from lower latitudes) – the tropical warming is greatest in the mid-to-upper troposphere because of changes in the moist adiabatic lapse rate (neutrally stable to moist convection); also, changes in surface evaporation and night-time cooling on land, etc…

    BUT I think Martin Vemeer (69) was not refering to the temperature response in a climate change, but rather to the spatial temperature variation at any one time.

    Temperature generally slows or stops decreasing with height at the tropopause and (by solar heating – ozone layer) eventually starts to increase until reaching the stratopause (in winter at higher latitudes, temperature continues to decrease with height above the tropopause (though less rapidly) for some distance; in the tropics, temperature increases nearly immediately above the tropopause, while elsewhere, the lower stratosphere can be nearly isothermal).

    Because of this, looking at the spectrum of LW radiation emitted to space, at wavelengths where there is great enough opacity, stratospheric emission dominates the emission to space. When going toward such opacity maxima, or when increasing the concentration of the relevant gas (including increases sufficiently high above the tropopause), the source of the radiation to space becomes concentrated higher up, and thus past the point where enough is coming from the stratosphere and not the troposphere or surface, the brightness increases. This does not contribute a negative radiative forcing at the tropopause, of course, because at such wavelengths, the tropopause level radiative forcing simply approaches zero as the source region for radiation in either direction becomes concentrated closer to the tropopause and thus has less temperature variation from one side to the other.

  32. 132
    Patrick 027 says:

    Just to be clear, I was eyeballing the annual cycle amplitudes from the graphs at – hence the ‘give or take’ note regarding South Pole values.

  33. 133
    Patrick 027 says:

    “A cross section is the effective area, normal to (facing) a particular direction, with which an amount of substance can absorb or emit as a perfect blackbody. ”

    Actually, though not of much importance to LW radiation (as opposed to solar radiation) under Earthly conditions, there is also the scattering cross section, which is the effective area that intercepts radiation without absorption but with redirection (scattering – reflection, refraction, diffraction). Scattering cross section + absorption cross section = extinction cross section.

  34. 134
    Martin Vermeer says:

    Steve R #115

    I think the 60% agrees reasonably with david’s model for water vapor increase when it says CO2 doubling direct sensitivity is 0.85 degrees for constant water vapor, and 1.26 degrees for constant relative humidity. What I did not find is why david’s model prediction is only 0.85 degrees for constant water vapor. Is that correct or not and why?

    Yes, I noted that too… And I would like to have the answer to that last question too. The value is wrong, and either the model computes wrong, or we’re all misunderstanding that “water vapor scale” thingy.

    We’re not the only ones exasperated: See last slide

    I don’t think David is actively following this blog, and Gavin has I think taken to not answering on others’ behalf. So let me try — nobody can mistake me for competent ;-) : this is a toy model. The atmospheric structure is assumed constant and equal to the US Standard Atmosphere, with lapse rate -6.5 degs/km, come rain or shine. The MODTRAN code behind it is a serious model (though a bit old) developed for serious work in military far-IR remote sensing. It was never meant to be used like this.

    What I did notice was that if you blow up “water vapor scale” to large values, the relative humidity in the data file output is restricted to 100%, even if the plotted curve is not. Wonder what the computation does.

    Jamie #128:

    Isn’t the resolution of ‘bands’ into ‘lines’ completely dependent on the scale at which one measures? Is there a single standard scale at which all electromagnetic spectra are universally measured? Does wavelength vary continuously, or in quantum steps? It seems to me that how many lines one finds must relate to the resolving power of the tool one uses. Is this wrong?

    Actually the limiting factor is a phenomenon called “line broadening”, merging the lines into bands. Quantum theoretically one can compute every line individually and its precise wave number, and for CO2 and H2O etc. there are thousands of them, associated with rotational and vibrational states, and they are very close together.

    There are two main line broadening effects: 1) doppler broadening, due to the molecules’ thermal motion, producing an exponential (gaussian) profile. And 2) pressure broadening, due to molecules being in the EM field of nearby molecules distorting their quantum states. This produces a Cauchy-Lorentz type profile, that sticks out over the doppler one for high enough pressures.

    Studied stellar atmospheres in a previous life… the Earth’s atmosphere isn’t so different ;-)

  35. 135
    Bill DeMott says:

    I’ve been reading fairly broadly on climate blogs. Here are two general fallacies that seem really common on denialists blogs. The first is that scientists have some how overlooked some very obvious factor, such as sun light, water vapor, the light absorbance characteristics of CO2, recycling of carbon in the biosphere etc. Any such statement suggests that scientists are completely incompetent. Scientists are constantly searching for some factor that has been overlooked or underestimated as a forcing factor. I can’t claim to have read extensively in the climate primary literature, but the notion that “low hanging fruit” have not been considered seems silly.

    The second fallicy is that some new scientific study proves that the whole basis of earlier scientific conclusions false. This hardly every happens in science and when it does, it usually takes many studies to change basic views and theories. This second fallacy is very similar to what one can read in the “anti-evolution literature.” Here minor progress (in understanding how evolution works in a particular case) is often taken completely out of context as proof that the “the theory is false.” These two fallacies seem to be bigger stumbling blocks for Americans, or at least for English-speaking counties, than for non-English speaking parts of Europe and Asia. I’m not sure why. Maybe it has something to do with distrust of authority and experts.

  36. 136
    Mark says:

    re #128

    FWHM is generally given as the definition of an absorbtion line.

    And that isn’t central to their rebuttal. Even if you consider it smeared over the 60,000 lines, it’s not those specific frequencies anyway. Which kind of is the point of “it’s only 8%” is not really the take-home message.

  37. 137
    Mark says:

    RodB: “Maya (72), in your answer to ziff you are terribly misreading or misrepresenting the “discrete” and narrow nature of the CO2 absorption bands.”

    Can you please tell us how narrow “narrow” is. You could be misrepresenting Maya’s comment: 60,000 narrow bands can (indeed MUST) cover quite a large area of the IR spectrum.

    Other queries to zif etc:

    1) 8% of the wavelength spectra or 8% of the energy released
    2) If 8% of the energy, at what temperature are you considering? 8% of the sun’s output isn’t in IR, and most of Deneb’s energy isn’t in the visible or lower frequencies. But my body radiates in a peak given by Wein’s displacement law. Which is happily within the IR band.
    3) 100% of the energy I’ve put into heating the water in my house will disappear unless I turn the heating on again, despite 97% of the heat loss being retained by the cladding. What importance does 8% have on heat retention and balance temperatures
    4) Please show your working on the 8%. After all, it’s easy to say it’s 28%. Who’s right? You with 8% or me with 28? How you worked it out is how you prove who is right.

  38. 138

    Mostly on-topic, here is a relatively upbeat assessment of the political and economic situation vis-a-vis mitigation in Asia. It underlines once again that the common Western perception that the large developing economies will have to be “dragged kicking and screaming,” so to speak, is less than entirely accurate.

  39. 139
    Maya says:

    “It’s common usage to refer to a band by its main but general wavelength, e.g. 15um (but not 15.2378um.)”

    Aah, now that’s a piece of information I was missing. Thank you.

    “a gasseous substance has individual absorption lines ….”

    Thank you, too, Patrick, for all of that. That helps a lot in understanding the context of what I’ve been reading on this, and I appreciate your effort in explaining.

  40. 140
    Maya says:

    Patrick – “Line strengths can also be temperature dependent.” Is that because of the doppler broadening effect itself? I’m picturing sort of a smearing effect, but I’m not sure if that’s a good way to think of it. If you draw a charcoal line on a piece of paper and then smudge it with your finger (metaphorically broadening the spectral line), the line changes from black to grey – the color (strength) is no longer as strong.

    Jamie, Martin (134) gave a better answer than I ever could in this lifetime! :)

  41. 141
    catman306 says:

    Off topic

    Please be aware of today’s (Mar 18) NPR Fresh Air: An interview with James Balog National Geographic photographer documenting the melting of glaciers with time lapse photography.

    Next week’s Nova National Geographic TV special will feature Balog’s work.

  42. 142
    Rob says:

    Does CO2 drive temperature rise or is it just UHI, according to the Jones et al study of 2008 40% of the increase in global temperature from 1951 to 2004 is from the Urban heat island effect.

    In it, Jones identifies an urban warming signal in China of 0.1 degrees C per decade. Or, if you prefer, 1 degree C per century.

    [Response: Read the paper a little more carefully. Jones et al suggest an urban effect in china (not globally) that reduces the regional trend (1950-2004) from ~1.3 deg C to 0.8 deg C. Still plenty of non-urban warming. Note too that this is with respect to nearby ocean temperatures which is not ideal. – gavin]

  43. 143
    Hank Roberts says:

    Check: given this claim, point to the “narrow bands of frequencies, which correspond to wavelengths of 2.7, 4.3 and 15 micrometers”

    ” An example of a terrestrial radiation spectrum measured at the top of the atmosphere by the Nimbus-3 IRIS instrument is shown in Figure 15.3. The absorbing bands such as the 9.6 µm band of O3 and the 15 µm band of CO2, as well as the atmospheric window and several other features (H2O, CH4), are noticeable.”

    So “narrow” is meaningless, and lacking any justirication for the eight percent claim, it seems pointless to try to figure out what the guy was talking about, eh?

  44. 144
    Patrick 027 says:

    Book (by Ray Pierrehumbert) with a lot of information about radiative energy transfer :

    Re 139/140 (you’re welcome :) ) – as I understand it, the line strength refers to all of the emission/absorption cross section per unit amount of substance attributable to that line. While conserving line strength, broadenning reduces the absorbance at and near the line center but increases it farther from the line center.

    The change in line strength due to temperature is a change in the emission/absorption cross section attributable to the line (PS changes in doppler broadenning and line strength due to temperature are not significant climate feedbacks, so far as I know).

    PS when adding up the cross section over a wavelength interval, it is not weighted by the radiation per unit wavelength (which varies over wavelength) – either for available radiation to absorb, or for blackbody emission – when finding a total cross section that is conserved by broadenning. OF course, one must weight by the radiant intensity per unit wavelength interval when calculating the effect of all this on radiative energy transfers.

  45. 145

    Hank Roberts wrote in 142:

    Check: given this claim, point to the “narrow bands of frequencies, which correspond to wavelengths of 2.7, 4.3 and 15 micrometers”

    After looking at that diagram, I got to thinking: for each wavelength, given:
    — the Planck curves that are labeled with their associated temperatures;
    — the lapse rate which tells you how quickly temperature drops relative to altitude; and,
    — the fact that local equilibrium conditions hold until above 20 mb of pressure,

    … one could calculate the altitude at which the atmosphere goes from being opaque to a given wavelength to being transparent to that wavelength by looking at which Planck curve the wavelength’s brightness temperature crosses.

    The diagram contains a lot of information.

  46. 146
    Chris says:

    Wait a minute Gavin. I can see that in Rob’s posted paper the data has only been collected and applied in China, but is that really grounds for completely dismissing it? I mean, Chinese cities aren’t fundamentally different from cities anywhere else, are they? The argument that Chinese cities have grown faster than others may be valid (I’m not sure if they really have), but the UHI effect was still underestimated by around a factor of 20 by the IPCC. This seems rather serious and deserving of discussion.

  47. 147
    snorbert zangox says:


    1. Physical Chemistry. a system of colloidal particles dispersed in a gas; smoke or fog.

    In other words, the cloud is an aerosol, the particles may be either liquid or solid.

    On to a more interesting subject. I just found an article that is appearing in Int.J.Mod.Phys.B23:275-364,2009, soon if not already. I have just now begun reading it but the abstract, below, is interesting to say the least.


    [Response: It’s nonsense. It was nonsense when it was first written two years ago, and it remains nonsense now. However it does stand as prima facie evidence that anything can make it into some obscure corner of the scientific literature if one is persistent enough. Thus the absence of serious contriarian literature becomes even more remarkable. – gavin]

  48. 148
    Patrick 027 says:

    Re 146 – perhaps you are well aware of this, but of course, there are ocean temperatures, receding glaciers, ecological changes, borehole measurements…

    If a station has been in an urban environment for some time, conceivably, changes in temperature would not be due so much to the urban heat island because it’s already there.

  49. 149
    Patrick 027 says:

    Re 145 – “one could calculate the altitude at which the atmosphere goes from being opaque to a given wavelength to being transparent to that wavelength by looking at which Planck curve the wavelength’s brightness temperature crosses.”

    Clarification: The source of radiation reaching any location, coming from any one direction, is distributed along a path; the brightness temperature at any wavelength is some effective temperature that is within the range of temperatures along the path.

    (clarification also for point in 144): The distribution of emission of radiation along a path that reaches a viewer is an exponential function when the path is measured in optical thickness; it is the same as the distribution of absorption of the radiation at the same wavelength in the opposite direction that is passing by the viewing point.

    The distribution exponentially decays away from the viewing point – proportional to exp(-optical thickness). The transmissivity along the path exponentially decays (from 1) along the path in the same way. The absorptivity and emissivity thus start at zero and assymptically approach 1 with increasing optical thickness.

    For sufficiently small optical thickness, absorptivity and emissivity are about equal to their emission cross sections per unit area facing the direction being considered; thus absorptivity and emissivity increase over geometric distance in proportion to cross section per unit volume. However, as the absorptivity/emissivity become significant, additional cross sections are partly blocked by the cross sections already added. Thus, optical thickness increases linearly with cross section per unit area (equal to cross section per unit volume multiplied by geometric path length), while absorptivity and emissivity ‘decay’ exponentially to 1 from zero.

    When scattering is significant, it get’s a bit more complicated; absorptivity and emissivity cannot reach 1; some of the radiation coming from a path will have been scattered into that direction… (PS a greenhouse effect can operate with LW scattering instead of or along with emission/absorption – this could happen in a ‘Snowball Mars’ wherein dry ice (CO2) cloud particles would scatter LW radiation. The mathematics will be a bit different but there are at least some qualitative similarities in so far as radiation is being blocked (but by backscattering rather than absorption – it would be like putting a double-sided mirror in between the surface and space; the atmosphere would control radiative transfer but not actually gain or lose energy directly by it if scattering occurs without absorption/emission).

  50. 150
    Patrick 027 says:

    “The distribution exponentially decays away from the viewing point ”

    So looking down from space, you can see a certain depth into the atmosphere (and some of the surface if that depth is great enough), not because the atmosphere is transparent above that level but because visibility of an object increases while visibility of what is behind it decreases gradually for increasing object thickness from zero, disregarding reflection at a surface (as is appropriate for bulk atmospheric properties).

    To a first approximation, where well-mixed optical agents dominate, one could assume optical depth is proportional to mass path – so in the vertical, optical depth as a vertical coordinate would be nearly proportional to pressure. At many wavelengths, however, optical thickness per unit mass per unit area decreases (up to a point, at least) with increasing height because of reduced pressure broadenning, while at a few wavelengths the opposite would occur. Because absorptivity/emissivity can only approach 1 (saturation), the concentration of cross section into smaller wavelength intervals results in more rapid saturation at some wavelengths along with reduced opacity at others, so that the overall opacity over a range of wavelengths, over some distance, is reduced by lack of broadenning.