Guest Post by Willis Eschenbach
My thanks to Nick Stokes and Joel Shore. In the comments to my post on the effects of atmospheric black carbon, Extremely Black Carbon, they brought up and we discussed the results of Ramanathan et al. (PDF, hereinafter R2008). Black carbon, aka fine soot, is an atmospheric pollutant that has been implicated in warming when it lands on snow. However, despite many claims to the contrary, atmospheric black carbon cools the surface rather than warming it.
There is an important implication in Ramanathan’s work regarding the canonical claim of AGW supporters that changes in surface temperature slavishly follow changes in forcing. Their claim is that the change in surface air temperature ( ∆T ) in degrees Celsius is a constant “lambda” ( λ ) called the “climate sensitivity” times the change in forcing ( ∆F ) in watts per square metre (W/m2). Or as an equation, the claim is that ∆T = λ ∆F, where lambda( λ ) is the climate sensitivity.
In R2008 they discuss the effect of black carbon (BC) on the atmosphere. Here’s the figure from R2008 that I want to talk about.
Figure 1. Figure 2C from R2008 ORIGINAL CAPTION: BC [black carbon] forcing obtained by running the Chung et al. analysis with and without BC. The forcing values are valid for the 2001–2003 period and have an uncertainty of ±50%. [Presumably 1 sigma uncertainty]
This figure shows the changes in forcing that R2008 says are occurring from black carbon forcing. Here is R2008’s comment on Figure 1, emphasis mine:
Unlike the greenhouse effect of CO2, which leads to a positive radiative forcing of the atmosphere and at the surface with moderate latitudinal gradients, black carbon has opposing effects of adding energy to the atmosphere and reducing it at the surface.
R2008 also says about black carbon (BC) that:
… as shown in Fig. 2, for BC, the surface forcing is negative whereas the TOA forcing is positive (Fig. 2c).
What are the mechanisms that lead to that re-partitioning of energy between the atmosphere and the surface?
Before I get to the mechanisms, I want to note something in passing. R2008 says that the forcing values have an uncertainty of ± 50%. That means the “Atmosphere” forcing is actually 2.6 ± 1.3 W/m2, and the “Surface” forcing is -1.7 ± 0.85 W/m2. This means that there is about a 30% chance that their “TOA” forcing, which is atmosphere plus surface, is actually less than zero … just sayin’, because Ramanathan didn’t mention that part. But for now, let’s use their figures.
PART I – What’s going on in Figure 1?
According to R2008, atmospheric black carbon causes the surface to cool and the atmosphere to warm. The surface is cooled by atmospheric black carbon through a couple of mechanisms. First, some of the sunlight headed for the surface is absorbed by the black carbon, so it doesn’t directly warm the surface. Second, any sunlight intercepted in the atmosphere does not have a greenhouse multiplier effect. Together, they say these effects cool the surface by -1.7 W/m2.
The atmosphere is warmed directly because it is intercepting more sunlight, with a net change of + 2.6 W/m2.
R2008 then notes that the net of the two forcings, 0.9 W/m2, is the change in the top-of-atmosphere (TOA) forcing.
The authors go on to say that because black carbon (BC) has opposite effects on the surface and atmosphere, the normal rules are suspended:
Because BC forcing results in a vertical redistribution of the solar forcing, a simple scaling of the forcing with the CO2 doubling climate sensitivity parameter may not be appropriate.
In other words, normally they would multiply forcing times sensitivity to give temperature change. In this case that would be 0.9 W/m2 times a sensitivity of 0.8 °C per W/m2 to give us an expected temperature rise of three-quarters of a degree. But they say we can’t do that here.
This exposes an underlying issue I want to point out. The current paradigm of climate is that the surface temperature is ruled by the forcing, so when the forcing goes up the surface temperature must, has to, is required, to go up as well. And vice versa. There is claimed to be a linear relationship between forcing and temperature.
Yet in this case, the TOA forcing is going up, but the surface forcing is going down. Why is that?
To describe that, let me use something I call the “greenhouse gain”. It is one way to measure the efficiency of the poorly-named “greenhouse” effect. In an electronic amplifier, the equivalent would be the gain between the input and output. For the greenhouse, the gain can be measured as the global average surface upwelling radiation (W/m2) divided by the global input, the average TOA incoming solar radiation (W/m2) after albedo. For the earth this is ~ 390W/m2 upwelling surface radiation, divided by the input of ~ 235 W/m2 after albedo, or about 1.66. That’s one way to measure the gain the surface of the earth is getting from the greenhouse effect.
Note that the surface temperature is exquisitely sensitive to the surface gain of the greenhouse effect. The gain is a measure of the efficiency of the entire greenhouse system. If the greenhouse gain goes down from 1.66 to 1.64, the surface radiation changes by ~ 4 W/m2 … on the order of the size of a doubling of CO2. Note also that the greenhouse gain depends in part on the albedo, since the 235W/m2 in the denominator is after albedo reflections.
Here is the core issue. For the “greenhouse” system to have its full effect, the sunlight absolutely must be absorbed by the surface. Only then does it get the surface temperature gain from the greenhouse, because some of the surface radiated energy is being returned to the surface. But if the solar energy is absorbed in the atmosphere, it doesn’t get that greenhouse gain.
So that is what is happening in Figure 1. The black carbon short-circuits the greenhouse effect, reducing the greenhouse thermal gain, and as a result, the atmosphere warms and the surface cools.
PART II – Almost Black Carbon
R2008 discusses the question of the 0.9 W/m2 of TOA forcing that is the net of the atmosphere warming and surface cooling. What I want to point out is that the 0.9 W/m2 of TOA forcing is not fixed. It depends on the exact qualities of the aerosol involved. Reflective aerosols, for example, cool both the atmosphere and the surface, by reflecting solar radiation back to space. Black carbon, on the other hand warms the atmosphere and cools the surface.
Consider a thought experiment. Suppose that instead of black carbon (BC), the atmosphere contained almost-black carbon (ABC). Almost-black carbon (ABC) is a fanciful substance which is identical to black carbon in every way except ABC reflects a bit more visible light. Perhaps ABC is what is now called “brown carbon”, maybe it’s some other aerosol that is slightly more reflective than black carbon.
As you might imagine, because almost-black carbon reflects some of the light that is absorbed by BC, the atmosphere doesn’t warm as much. The surface cooling is identical, but the almost black carbon reflects some of the energy instead of absorbing it as black carbon would do. As a result, let us say that conditions are such that ABC warms the atmosphere by 1.7 W/m2 and cools the surface by -1.7 W/m2. There is no physical reason that this could not be the case, as aerosols have a wide range of reflectivity.
And of course, at that point we have no change in the TOA radiation, but despite that the surface is cooling.
Which brings me at last to the point of this post. To remind everyone, the canonical equation says that the change in surface air temperature ( ∆T ) in degrees Celsius is some constant “lambda” ( λ ) times the change in TOA forcing ( ∆F ) in watts per square metre (W/m2). Or as an equation, ∆T = λ ∆F, where lambda( λ) is the climate sensitivity.
But in fact, all that has to happen to make that equation fall apart is for something to interfere with the greenhouse gain. If the efficiency of the greenhouse system is reduced in any one of a number of ways, by black carbon in the atmosphere or increase in cloud albedo or any other mechanism, the surface temperature goes down … REGARDLESS OF WHAT HAPPENS WITH TOA FORCING.
This means that the surface temperature is not simply a function of the TOA forcing, and this clearly falsifies the canonical equation.
In fact, I can think of several ways that surface temperature can be decoupled from forcing, and I’m sure there are more.
The first one is what we’ve just been discussing. If anything changes the greenhouse thermal gain up or down, the TOA radiation can stay unchanged while the surface radiation (and thus surface temperature) goes either up or down.
The second is that clouds can decrease the amount of incoming energy. It only takes a trivial change in the clouds to completely counterbalance a doubling of CO2. This is a major function of the tropical clouds, which counteract increasing forcing by forming both earlier and thicker.
The third is that the system can change the partitioning between the throughput and the turbulence. The throughput is the amount of energy that is simply transported from the equator to the poles and rejected back to space. On the other hand, the turbulence is the energy that ultimately goes into heating the climate system. In accordance with the Constructal Law, the system is constantly evolving to maximize the total of these two.
Fourth, the El Nino/La Nina system regulates the amount of cool ocean water that is brought to the surface, as well as increasing the heat loss, to avoid overheating. (One curious consequence of this is that the surface temperature in the El Nino 3.4 area has not warmed over the entire period of record … but I digress).
Part III – CONCLUSIONS
The conclusion is that the simplistic paradigm of a linear relationship between temperature and forcing can’t survive the observations of Ramanathan regarding black carbon. For the surface temperature to vary without changes in the TOA forcing, all that needs to happen is for the greenhouse thermal gain to change.
w.
APPENDIX- How it works out
For the math involved, let me steal a diagram from my post, “The Steel Greenhouse”
Figure 2. Single-shell (“two-layer”) greenhouse system, including various losses. S is the sun, E is the Earth, and G is the atmospheric greenhouse shell around the Earth. The height of the shell is greatly exaggerated; in reality the shell is so close to the Earth that they have about the same area, and thus the small difference in area can be neglected. Fig. 2(a) shows a perfect greenhouse. W is the total watts/m2 available to the greenhouse system after albedo. Fig. 2(b) is the same as Fig. 2(a) plus radiation losses Lr which pass through the atmosphere, and albedo losses ( L_albedo ), shown as W0-W. Fig. 2(c) is the same as Fig. 2(b), plus the effect of absorption losses La. Fig. 2(d) is the same as Fig. 2(c), plus the effect of thermal losses Lt. These thermal losses can be further subdivided into sensible ( L_sensible ) and latent heat ( L_latent ) losses (not shown).
We are interested in panel (d) at the lower right of Figure 2. It shows the energy balances.
As defined above, the thermal gain ( G ) of a greenhouse is the surface temperature (expressed as the equivalent blackbody radiation) divided by the incoming solar radiation after albedo. In terms of the various losses shown in Figure 2, this means that the greenhouse thermal gain G is therefore:
where
The important thing to note here is that if any of these losses change, the greenhouse gain changes. In turn, the surface temperature changes … and the TOA balance doesn’t have to change for that to happen.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
rgbatduke says:
Your two posts are really good. Short catalogues of potential research projects, so to speak. You probably will like Isaac Held’s blog: http://www.gfdl.noaa.gov/blog/isaac-held/2011/10/26/19-radiative-convective-equilibrium/
If you already know of those, then I apologize for the redundancy.
One can imagine how such simulations might be inhanced/complicated by adding variaties of ABC effects, as well as how hard that would be to do in practice. Hard, but about the right level of difficulty for a next step, IMHO. All throughout atmospheric science, going from the approximate first-order effects as Willis did here, to detailed studies of processes, entails huge increases in the efforts required.
Wills and EM Smith
I recently saw a post http://perspectives.mvdirona.com/2012/03/17/ILoveSolarPowerBut.aspx by James Hamilton- an IT guy- about the overall effectiveness of PV. Your post and discussion got me to thinking a bit about albedo (Willis’s- ” Note also that the greenhouse gain depends in part on the albedo, since the 235W/m2 in the denominator is after albedo reflections.” ). Specifically James notes: “Let’s focus instead on large datacenters in rural areas where the space can be found. Apple is reported to have cleared trees off of 171 acres of land in order to provide photo voltaic power for 4% of their overall estimate data center consumption. Is that gain worth clearing and consuming 171 acres?”
I am trying to figure out how answer James’s question above. It seems like one should try to include the change in albedo in any evaluation of benefits vs. costs. EM comments about grid cells leads me to this conclusion as well. I think I have a basic understanding on how forcing is required in the climate models vs. my understanding of how the physics, chemistry and biology seems to work at my little ranch/farm. What I am interested in knowing currently is if the climate models allow for the albedo to change with what we humans (or the natural environment) do in response to changes in the incoming energy from the sun or the forcings?
In any case, Thanks for the learning experience! It seems like one should try to include the change in albedo in any evaluation. A recent summary of Severin Borenstein seem important to consider too when looking at the effectiveness of our activities- http://ei.haas.berkeley.edu/pdf/newsletter/2012Spring.pdf
“Likewise, the pollution benefits from renewable energy depend on what type of generation it displaces, which also depends on time and location. Without incorporating these factors, cost-benefit analyses of the alternatives are bound to be misleading. If governments are to implement reasoned renewable generation policy, it will be critical to understand the costs and benefits of these technologies in the context of the electricity systems in which they operate.”…………..
This is very interesting, but the main issue usually discussed about black carbon is it causes melting of the ice and snow. It still does this. Soot landing on ice does not, of course, cause the surface to warm up per se but rather causes the ice to melt as the carbon heats up – (presumably if the ice is not too far below freezing).
1) How come that as Carbon Dioxide (CO2) warms up as it receives “Thermal Radiation” (TR) from the Surface, it (CO2) can send the radiation it has gained back down to whence it came, but as the now Atmospheric Black Carbon (ABC) absorbs TR from the Sun it can not direct its TR towards the surface thus offset any cooling.
meters per second (which would usually cross that distance without a collision in 0.027 milliseconds, or roughly 170 times as long). Even this estimate isn’t right — illustrating the problem with doing this any simple way — because the mean free path of the photons gets longer the higher the photon diffuses through the bumpers, so it goes farther and has a greater chance of escaping. In reality the transit is probably somewhere in between the 0.026 milliseconds and 4.5 milliseconds; I’d guestimate it around 1 millisecond but really it should be computed, as should the fraction of return pathways.
All CO_2 does is absorb and reradiate energy in its (fairly wide composite) absorption band(s) in the IR part of the spectrum. As electromagnetic radiation passes a CO_2 molecule, there is a finite probability that the molecule will absorb the radiation (exciting one of its quantum states) and then reradiate it. The catch is that it is likely to reradiate it in a somewhat random direction.
If you’ve ever played pinball, you can then visualize what the world looks like to an “IR photon” (easier to visualize in play than a classical EM wave). Some molecule down on the Earth’s surface kicks it into play just like the spring launcher on a pinball table, headed towards outer space. But noooo, there are too many bumpers in the way! It hits one half a klick up that sends it off at right angles. Now it is heading towards outer space the long way, no chance to escape. But bang! It hits another bumper 300 meters further on! This one deflects it back up again, but now it is heading up and east. 1200 meters in that direction pow, only this time it is deflected back towards the ground.
Maybe it makes it! When it hits the ground it is absorbed, briefly warming the ground where it hits, but then the ground cools by kicking a new photon into play heading at the sky. This one goes 2 klicks before hitting the first bumper, then 1 klick south, then 500 meters up, then 900 meters west, then 2700 meters northwest and up, then… maybe it makes it out the far side headed towards space. Maybe it diffuses by means of many collisions back down to the Earth somewhere else. In any event it now takes far longer (on average) for the humble IR photons in the CO_2 absorption bands to make it out of the Earth’s atmosphere, and a very significant fraction of them return to the Earth’s surface for at least one more visit there (sustaining its emission temperature) along the way.
For what it is worth, the mean free path of our CO_2-resonant photon is actually only around 47 meters — the numbers above were illustrative only. That means that on average the photon won’t make it across half a football field before being scattered. Even travelling at the speed of light (and allowing nothing for the photon is absorbed “in” the CO_2 “bumpers”) it takes 4.5 milliseconds to cross the 8 kilometers of atmosphere for a photon that moves at
This mental picture is important because:
* It is the right mental picture for “downwelling radiation”. of the sky as a vast vertical three dimensional pinball machine with every point on the ground launching a steady stream of pinballs straight up (and no “gravity). Those pinballs enter bumpersville and diffusively bounce around until they either exit at the bottom, to be queued up and fired again, or finally happen to hit a clear pathway to space at the top. At any instant in time, there are far more photon/pinballs in transit (order of 100x as many!) as there are being fired in and/or emerging on the far side because they are being lagged order of 100x the straight-out transit time. Plenty of pinballs are shot straight up only to recoil straight back down at the ground.
“Ground temperature”, BTW, is basically the height of the pile of pinballs that the ground has to fire. Every time a pinball (fired anywhere) bounced back to ground it has to go on the stack for that particular point and be fired again, and in the meantime the stack stays a bit higher than it would if no balls ever returned! Greenhouse warming!
* This is the same general model for the albedo of clouds, why clouds (or aerosols, or black carbon dust) reduce the passage of light visible or invisible through them. Small particulates act like pinball bumpers, “trapping” radiation in caged motion that ultimately rejects a significant fraction of them back the way they came (and thereby reduce the average transmission in the original direction.
* In the case of an open system, where a steady stream of pinballs is being delivered that do not scatter off of the intermediary bumpers (coming in as visible light) but that are transformed on hitting the ground into IR pinballs that do scatter off of the bumpers, where the “fire rate” of the ground launchers is proportional to the height of the stack to the fourth power, the ground quickly steps up its fire rate to compensate until ins (on average) equal outs, where “average” isn’t at all fine grained — most of the time every single point on the Earth’s surface is in imbalance, either net warming or net cooling.
* Just for grins, the mean free path for a photon in the sun is small enough that the diffusion time for a photon produced in the solar interior is (IIRC) around 100,000 years. This creates a hellish thermal differential “greenhouse effect” that helps keep the core hot enough to sustain fusion.
Even so, this is still incomplete and inadequate. Humid air has water molecules that are far more prevalent and active and reduce the mean free path to order or 8 meters! Also, if the CO_2 molecule collides with an air molecule during the time it has absorbed the photon but before it emits it, it can (and will) “heat” the air molecule (transfer energy to it), effectively removing the IR pinball from play as the air radiatively cools entirely differently. Finally, a lot of the energy absorbed in this way (warming of the air) is convected to the top of the troposphere where it transfers back to the CO_2 and is radiated away — there is substantial bulk transport of heat, not just radiative transfer, within the atmosphere. Clouds (present or absent), particulates — everything changes, to where my simple “pinball” metaphor breaks down or becomes as hopelessly complex as the physics.
But this should answer, at least in one easily understandable metaphorical way, your (frequently asked) question. Think of photons as pinballs and CO_2 as bumpers that are dense enough that the pinball can go only 1/200th of the way to the top of the troposphere (where it has a good chance to escape out of play) before hitting one, no matter what direction it goes. That makes it easy to see why photons have a lot harder time making it out, and why many of them are scattered back to the ground before they do (basically reducing the rate of energy transfer out of the ground, the cooling rate, until the ground heats up enough to compensate and keep rates in balance anyway.
rgb
One can imagine how such simulations might be enhanced/complicated by adding varieties of ABC effects, as well as how hard that would be to do in practice. Hard, but about the right level of difficulty for a next step, IMHO. All throughout atmospheric science, going from the approximate first-order effects as Willis did here, to detailed studies of processes, entails huge increases in the efforts required.
I agree. Indeed, my personal area of physics expertise is simulations (that’s why I’m so cynical about GCMs, which are gross differential first order models, in a lot of cases, where the underlying reality is nonlinear, partial differential, stochastic diffusive, and has more control variables than they think that it does.
Like leaving out the albedo change. I mean, why bother publishing it if nobody in climate research are going to respond to this huge news? Albedo isn’t like an adjustment on your deductions on a complex schedule where it is hard to see if you win or lose, it is like changing your tax rate. It directly varies the temperature, far more than any “forcing” change. 2% of the total TOA insolation is enormous, and that’s what changing 0.3 to 0.32 represents. True, it is shrunk a bit by the fourth root in the BB formula, but it still vastly outweighs trivial modulation of the WORST CASE direct CO_2 based warming. Which is why unless it goes back up, it will get much cooler — cooler by degrees C — over the next decade or so as the oceans slowly shed heat. If the albedo is indeed connected to solar magnetic activity, and we are indeed entering a Maunder minimum (next cycle and perhaps the one beyond) then it could be 2050 or later before the albedo returns to “normal”; which will be far less than its 20th century Grand Solar Maximum minimum.
rgb
These discussions seem to me to be among the most important I’ve read. I particularly appreciated the various explanations of ‘forcing’, which was a concept appearing in no physics course or text in my experience, and is therefore one to be treated with a great deal of scepticism.
rgbatduke says:
March 28, 2012 at 2:35 pm
Well written.
One question about this: Also, if the CO_2 molecule collides with an air molecule during the time it has absorbed the photon but before it emits it, it can (and will) “heat” the air molecule (transfer energy to it), effectively removing the IR pinball from play as the air radiatively cools entirely differently.
When this occurs, energy in the excited electrons is transmitted to the air molecule? Right?
rgbatduke says:
March 28, 2012 at 9:23 am
“the Earth almost certainly responds to any increase in “forcing” by also increasing the rate that it sheds heat. That is, net feedback to any forcing is almost certainly negative because positive feedback is associated with critical instability and there is no evidence of positive critical instability in the Earth’s climate record, recent or prehistoric…it would be a lot better to have models that exhibit this negative feedback as a direct outcome of self-organization within the system such that the system opposes externally driven changes everywhere except near critical points.”
Fascinating posts. Presumably, youe quote above offers a good explanation why the Earth’s climate system doesn’t (cannot) respond effectively to the onset of (Milankovitch) ice ages…. And, moreover, it tells us that the real climate risks humanity faces in the (relatively) distant future are: 1) from precipitous cooling (ice age onset) 2) are entirely natural (not man-made) in origin, and 3) we would need to find a way to reversibly (emphasis on that word) override the negative feedbacks in the climate thermoregulator at precisely the right time (or move to the tropics)…?
“rgbatduke says:
March 28, 2012 at 9:23 am
The thing that I think deserves the most attention is self-organization a la Prigogene. For reasons that I cannot fathom, I never hear of the world’s climate system described as a self-organized driven thermodynamic system far from equilibrium, in spite of the fact that that is precisely what it is.”
rgbatduke, do a search for Maximum Entropy Production (MEP) in non-linear / non-equilibrium fluid systems. Many papers on earth climate under MEP by Axel Kleidon, Ralph D. Lorenz and GW Paltridge, like :Entropy Production by Earth System Processes
Also used in studies on other planets and regarding Ice Ages.
OK, so no answers to my questions.
As I understand it, in order for Carbon dioxide to absorb and re-radiate IR wavelength energy, the high frequency visible and UV sunlight has to impact a surface and then be re-radiated as lower frequency IR.
Whilst in a black box conceptual model all of the incident heat will be re-radiated, the Earth is not a black box. Rather it is a complex mix of materials with all sort of properties that affect heat transfer.
Remembering that in such an environment, heat will be transferred by three modes (not one) Radiative, conductive and reflective, the exact amount of heat that will be radiated, depends majorly on the materials that receive the incident sunlight and since two thirds (?) of the planet is liquid this represents one very long term heat store.
True that ultimately, for the planets surface to be in thermal balance all incident heat need to be radiated back into space, but in a complex system, this radiation can take some time and some pretty long pathways.
All of which ultimately leads to weather and climate and since some of the climate cycles are pretty long then it can be hard to be certain what the heat balance is like given a snapshot. Which is what the whole debate is about in the first place.
So as far as black carbon is concerned if it absorbs heat and re-radiates it and assuming it is in the troposphere then I can’t see why this would change much at all. Except that ultimately they will be precipitated onto the surface and some will affect the albedo of snow and ice – much will not, because they will be washed out in rainfall.
As far as aerosols are concerned they clearly have some reflective (albedo) properties which means that they will reduce the incident sunlight on the planets surface, but if there is back carbon about some of that incident sunlight might neverthless be absorbed by these particles.
So the major aerosols are sulphate and nitrate, both of which are soluble so would be removed in precipitation, but are unlikely to darken surfaces, just change the pH of the precipitation. These aerosols can also be removed by dry precipitation with the surface and naturally can be absorbed/react in the oceans and waterways.
Remember that the “brown smog” that we see is not really brown, this is caused by backscatter of certain wavelenghts of sunlight by particles, in much the same way as Rayleigh Scattering in the atmosphere causes blue wavelengths to reach the planets surface – hence we see a blue sky during the day, unless you live in LA where looking directly up at midday gives a white haze, but a brown haze when the sun is lower in the sky or when you are looking across from the mountains.
Sorry, just to add a little more to my comments – White haze means that most wavelenghts of visible light is being reflected back, brown haze merely means that only certain visible wavelengths are being back scattered (reflected). This says very little about UV and whether those frequencies can penetrate aerosols. My guess is that they can to some extent in the same way as they can to some extent when its cloudy. Clouds are not perfect reflectors by the way.
A general comment on the discourses of this thread:
A widely used and well accepted but misleading terminology is “reradiate” or “re-radiate.” Typical statement such as “carbon dioxide absorbs IR then reradiates the energy,” implies that carbon dioxide does not emit IR if it does not absorb first. Radiation physics, as indicated by the SB equation, says that CO2 emits 24/7 as long as its temperature is not 0 K, regardless whether it absorbs or not.
We now have a multiple choice question: Are atmospheric CO2 molecules
1) warmer, 2) cooler or 3) the same than N2 and O2 molecules ?
I wonder how people will answer this question.
Mmm good point though a little pedantic – the point is that CO2 adsorbs infra-red at wavelengths that are within the range of IR wavelengths emitted by the Earth (which must be surface temperature dependent) whilst N2 and O2 molecules presumably don’t (otherwise they’d be greenhouse gases too). That being the case the term re-radiate doesn’t seem too innapropriate. As to the next question – it probably depends on whether the atmosphere is at the same temperature as the Earth’s surface beneath it.
My question back is how does heat get transferred to other molecules if they don’t absorb infrared and gases are not good conductors.
Goldie;
It is a key conception that many have not captured. We all know absorption leads to warming of an object. But this is in reference of before and after absorption of the object, not in reference of an absorbing object with non-radiative objects. For CO2, the temperature before absorption is 0 K, absorption of the Earth radiation warms it to around 200 K.
N2 and O2 do not absorb, so they do not emit. Therefore N2 and O2 will keep the heat they have without losing any. They pass their heat by molecular collisions with CO2 and H2O, which then emit the heat into space. Indeed, CO2 is a cooling agent for the atmosphere. Without so called greenhouse gases, the atmosphere would be far warmer and have a positive lapse rate, as we have observed in the high altitude thermosphere – where CO2 has been souted out due to heavier molecular weight.
When this occurs, energy in the excited electrons is transmitted to the air molecule? Right?
This is one of several things that can occur. When the molecule absorbs the photon, which carries momentum and energy, it recoils. It then holds the energy for a characteristic (exponential decay) time typically on the order of nanoseconds (but highly variable) and re-emits the photon at or near the original frequency (or a cascade of photons at lower frequencies, depending on the coupling of the excited level to lower levels) returning eventually to the ground state. The re-emitted photon is correlated in terms of polarization and emission direction with the original photon, but the original photon is on average unpolarized and on average the new emission direction and polarization are random. As the photon is re-emitted, the atom recoils again, this time in a new direction. The frequencies of the absorbed and emitted photons are tuned and detuned relative to the natural frequency of the level involved by a small amount due to the Doppler shift, a phenomenon called “inhomogeneous broadening”. Collisions that occur introduce random shifts in phase (minimally) and sometimes energy that further alter the frequencies (homogeneous broadening). The observed emission (and absorption) spectrum is thus not infinitely sharp for any given pair of coupled quantum levels, and the line widths do depend on the temperature (inhomogeneous) and pressure (homogeneous).
One thing I don’t understand is the assertion that increasing the CO_2 concentration will somehow alter the size of the CO_2 window by some sort of line broadening mechanism. Homogeneous broadening doesn’t depend on the partial pressure of CO_2, it depends on the mean free time between collisions, which depends on the total pressure and temperature of the atmosphere. It doesn’t matter what molecule it collides with, only that it collides. Similar Doppler broadening doesn’t depend on partial pressure.
In any event yes, when a collision occurs some energy and momentum is transferred from the CO_2 molecule to whatever it collides with. The collision also alters the local field and blurs/mixes quantum levels on both participants in the collision and hence enables electronic energy transfer, but in many/most cases this will be minimal and only have the effect of broadening the CO_2 emission lines (and facilitating absorbing in one level pair but emitting in one or more different ones in the resonant band of nearby levels). The direct transfer maintains thermal equilibrium between the CO_2 and the rest of the gas so they aren’t at different temperatures, so that the “warming” of the CO_2 due to absorption and “cooling” due to emission are shared and it doesn’t become seriously disequilibrated with its surroundings.
Increasing the concentration of CO_2 will reduce the mean free time and path between “bumper collisions” as the photons zig and zag diffusively (on average) outwards. This does in turn increase the “resistance” of the atmosphere to outflowing radiative energy in the relevant bands. It is by no means a linear increase — for one thing, the mean free path cubed is proportional to the volume per molecule, so doubling concentration only reduces the mean free path by around 20%. For another, the mean free path increases rapidly with height. For a third, CO_2 is only one greenhouse gas, and the other — water — is a comparative elephant to the CO_2 hare, with a mean free path a fifth as large already and with specular diffusion through and reflection from big puffy clouds to contend with as well. Finally, the atmosphere is already optically quite thick with respect to CO_2, and the outgoing radiative temperature in the CO_2 band has more to do with the height at which the atmosphere becomes “transparent” to in-band IR, e.g. the mean free path gets large enough that further collisions with bumpers become unlikely for photons emitted in the upward direction. Statistically, photons that diffuse to any given height preferentially diffuse upward at an increasing rate because the atmosphere below them becomes more reflective (on average) than that below them. All of these conspire to make the variation of temperature with CO_2 concentration rather weak once optical thickness is achieved, as it has been a hundred times over. Lifting the radiative height (and hence modulating radiative cooling in the IR band) seems to be more a function of the decadal oscillations and uplifting circulation (it is a big factor in El Nino, for example) than it is CO_2 concentration per se.
A really interesting subject, actually. As I have repeatedly said, a sound skeptical position is not “there is no such thing as the CO_2 based GHE”, as one can directly photograph it in action in TOA IR spectroscopy and fully understand its general mechanism, as I’ve tried to explain. It is its egregious amplification in models that try to include its effect by neglecting other important variables and modulators (enough to be able to explain historical natural variability, for starters) followed by an arbitrary multiplication by a non-observable “sensitivity” that is nothing less than a political fudge factor that will come out to be whatever you want it to be in a model you build so that it does that we should (and usually do) all object to.
All the reliable post 1980 data suggests that climate sensitivity has an upper bound that is slightly less than the lower bound of the earlier IPCC estimates. It is time that these estimates were fixed. If the climate stubbornly persists in cooling or remaining neutral, they will have to be fixed no matter how passionate the political inclinations of the scientist involved. One can only hide the lack of a rise and deviation from predictions based on high sensitivity for so long. It is almost certain, in my opinion, that the sensitivity is actually negative, that the real cause of the late 20th century warming was the grand solar maximum and albedo minimum that occurred then, and that when (eventually) the models are corrected to account for that source of natural, non-anthropogenic variability is accounted for the anthropogenic contribution to warming from the 20th century will end up being order of or less than 0.5C including all feedbacks! This is utterly negligible, and would continue to be negligible extrapolated to an end of the twenty-first century doubling of CO_2, assuming that any warming caused isn’t completely erased by the natural downward variation likely to be associated with the recent increase in bond albedo back to pre-20th century levels and beyond.
rgb
rgb;
What a superb series of posts. Your explanations are much appreciated. You have made explicit and clear many of the positions I have been stumbling towards. I get ignored or hammered every time I insist that “no-feedback sensitivity” is tiny compared to Hokey Team and IPCC estimates, well below 1°. Thank you for justifying my stubbornness!
Not sure if WE will take on board your analysis showing that BC and ABC increase airborne albedo and reduce melt-season snow-cover ground albedo. But it’s worth a shot!
What are the consensual atmospherics at Duke like? Are you a shunned outlier, or representative?
RE:
Hans says:
@ur momisugly March 28, 2012 at 6:09 pm
Many thanks for the link etc…
Andrew, here two more links explaining the MEP mechanism working in our climate,
The
second law of thermodynamics and the global climate system: A review of the maximum
entropy production principle
Nonequilibrium
thermodynamics and maximum entropy production in the Earth system
The Second Law still rules, have fun.
rgbatduke says:
March 29, 2012 at 9:59 am
You wrote a better answer than my question deserved. I was thinking along the lines of “Absorption raises the electrons to higher energy states; emission occurs when they return to the lower state; if there is a collision then the high energy electrons are reduced to lower energy states without emission of photons .” I didn’t know if it was possible, and my ability to formulate a question totally vanished.
Many thanks.
rgbatduke says:
March 28, 2012 at 5:09 am
Duke, I couldn’t agree more. I merely used Ramanathan’s findings to make a point about the variability of the response of the atmosphere to aerosols, not because I particularly believed them. That’s why I pointed out the error in the forcings of ± 50%.
Ramanathan’s results are important because in so many papers the claim is made that if forcing changes, surface temperature must change. The R2008 findings show this to be false, they show that TOA forcing can increase while surface temperature drops. And that is the sole reason I discussed them, because I have the same considerations you have about the R2008 findings.
My rule is, nature simply isn’t that simple. The height, the size of the particles, all of that matters.
By the way, there is a model that can actually resolve some of this stuff. It’s the brainchild of one man, Mark Jacobson, as good things often are. It’s called GATOR-GCMOM, and it does the stuff the other models only pretend to do.
Yes, at the end of the day it’s still a model … but it’s a reasonable one.
All the best, thanks for the comments,
w.
RE: Hans says:
@ur momisugly March 29, 2012 at 6:51 pm
———————
Many thanks Hans. I really enjoyed the Introd chapter you linked to (Kleidon & Lorenz) and similarly look forward to reading the papers you subsequently linked to (above). Because I haven’t yet read these, my questions/ points below might either or both be moot and/or uninformed:
Q1 (request really): In a post above you noted that the ice age phenomenon is addressed by the MEP principle in a paper/ chapter. Grateful if you would point me at it if it is not one of the two later links you provided…
Q2. the main problems I have with the ‘Gaia hypothesis’, apart from its unfalsifiability (and thus it does not really qualify as a hypothesis) are two-fold:
1. seems to me the conceptual power of complexity theory is that it provides a paradigm whereby the Earth’s biota can be regarded as an emergent property of the physical universe – not the other way round. Climate stability, as an emergent property of physcs, was obviously necessary before life (biological complexity) could arise (self-organise). Sure, the biota then modifies the atmosphere but the physics necessary to generate complex climates which generate temperature stability must have first been satisifed.
The term Gaia, however, in its varied fuzzy interpretations seems to imply this is the other way round doesn’t it? What additional explanatory value does the Gaia hypothesis provide to an understanding of how complex physical systems such as the Earth’s climate arise and operate? Certainly, prior to the formation of life, none – whereas the thermoregulating behaviour of the Earth’s climate system was clearly pre-requisite to the emergence of life on this planet…
2. The politicisation of climate science seems to me to have been deeply wound up with this ill-defined and untestable term ‘Gaia’. If it wasn’t so fuzzy, if it was a real hypothesis (ie. testable) there would be less cause for concern and the hysteria around CAGW could have been stemmed long ago. There’s a danger isn’t there in fuzzy terms masquerading as scientific hypotheses?
What are your thoughts?
Fascinating posts. Presumably, your quote above offers a good explanation why the Earth’s climate system doesn’t (cannot) respond effectively to the onset of (Milankovitch) ice ages…. And, moreover, it tells us that the real climate risks humanity faces in the (relatively) distant future are: 1) from precipitous cooling (ice age onset) 2) are entirely natural (not man-made) in origin, and 3) we would need to find a way to reversibly (emphasis on that word) override the negative feedbacks in the climate thermoregulator at precisely the right time (or move to the tropics)…?
Ice ages are interesting. The Earth’s climate is currently at least bistable — “cold phase” (glacial, ice age) and “warm phase” (interglacial). The cold phase is more stable, evidenced by spending (very) roughly 90,000 years in cold phase vs 10,000 years in warm phase over the last five or more cycles. I have studied bistable open systems in the context of quantum optics (and have a long Physical Review paper on the subject that presents the results of a microscopic simulation from way back in the 80s) as well as magnetic bistability and critical dynamics, so I do have a pretty good grasp on the kinds of differential systems that can lead to it.
The way that it typically works is that the system exhibits hysteresis (look it up on Wikipedia if need be). Depending on where one is in the parameter space that drives the transition, there may be either only one stable state for the climate — either warm phase or cold phase, only a single stable solution and perturbations will always drive one back to the vicinity of that solution — or there may be two stable solutions (or even more than two, especially in openly chaotic systems which may have many attractors and be multistable, not just bistable).
Where there are two solutions, the one you find yourself in is locally stable. If you perturb the system by “small” amounts, warming or cooling in the case of the Earth’s climate system, it will generally return to the locally stable phase. However, the other phase is also locally stable, and if you perturb it too far in that direction, you will cross a critical line that makes the other phase the locally stable state and rapidly switch to that state whether or not the perturbation ends.
To be explicit, ice has a very high albedo and takes a lot of heat to melt. If one covered the Earth today with (say) ice all the way down to the line marking the maximum glaciation in the last glacial period, it would reflect lots of sunlight without giving it an opportunity to heat the Earth. If the mean albedo increased to (say) 0.4, that would be enough to drop the Earth’s average temperature roughly 10 degrees Kelvin, which in turn might make it cold enough to sustain that ice instead of gradually melting it. If it did, the ice age would persist until something changed to favor gradual melting, e.g. a reduction of the albedo (which would then feed back by melting the ice to reduce it still further) or some other alteration that “raised the local thermostat” to where warm phase was the only stable state.
We have no idea what the parametric boundary is for that sort of bistable transition, but what we do know is that as the unknown underlying parameters that eventually will make the warm phase completely unstable and force the Earth into a completely stable cold phase move in that direction, the system will be bistable and furthermore will have a local stability boundary that gradually gets closer and closer to the stable warm phase norm.
This has two observable effects. One is that it will take smaller perturbations to flip the Earth into cold phase. Because the boundary is a place where the system isn’t strongly driven either way, as perturbations move the Earth towards that boundary, it will spend more and more time there before returning to “normal” temperatures for warm phase. The other is that because the stabilizing “forces” are relatively weakening in the warm phase, all natural fluctuations will tend to have longer lifetimes and greater excursion — the “noise” in the climate system will increase and it will become less predictable. All of these sorts of things are observed in naturally bistable physical systems as they near a critical point (in this case a “first order” critical point).
In that regard, bearing in mind that the timescale of fluctuations is at least decades to centuries for the Earth’s climate system, there are two or three data that should be worrisome to us. First is the LIA. It was the coldest excursion in the entire Holocene, post the Younger Dryas (a cold phase bistable fluctuation that actually drove us back to the disappearing cold phase briefly after a fluctuation had kicked us briefly out of cold phase (but obviously not stably) at the start of the Holocene. This could be because of extreme circumstances in the drivers — the Maunder Minimum — but even a Maunder Minimum might have have created such a large excursion if we were solidly and stably in warm phase. The second is the large and fairly rapid variability of the climate observed over the last 1300 years. Earth has gone from the warm MWP through to “normal” Holocene temperatures (but with fairly large fluctuations, dropped to a 10,000 year minimum and stayed there for close to a century with glacial growth, suggesting that for a while there cold phase was emerging as a stable attractor, then warming briefly to normal to chill again in the Dalton minimum, then warming steadily back to normal and beyond, all in the space of some 400 years.
A large part of this movement has probably been driven by solar state, but we don’t know what actually controls the emergence of warm/cold phase bistability and we are completely clueless about the shape of the stability boundary across the bistable regime (which is probably a bistable surface over a multidimensional space — or worse). You invoke Milankovitch, but this is not a sufficient explanation for the bistability observed. It has the wrong periodicity, for example. It is probably a factor in this multidimensional space, but not the factor.
In that regard, CO_2 may have been rather fortunate — if cold phase is emerging (and of course, the empirical evidence is that it is emergent, although we don’t know when — cold phase stability may already exist and the boundary may already be creeping north towards warm phase temperatures) then warming the Earth, even a little bit, may be stabilizing warm phase, protecting it from potentially critical fluctuations that might flip us to cold phase. It may also have reduced the climate sensitivity (by sharpening the restoring forces that make warm phase stable in the first place. There is a bit of empirical evidence that this is the case (if anything, the frequency and violence of storms and prevalence of drought appears to have diminished a bit compared to very long term data — e.g. the east coast drought in the 1600s that nearly wiped out all the European colonies and a number of indigenous tribes down throughout the southeast). It could also be why the IPCC estimates are wrong — they may be basing their sensitivity assumptions on the variability post LIA (the blade of the hockey stick) without recognizing that this may have been restabilizing the system in warm phase where it had begun to be unstable, so that sensitivity is now actually reduced.
Sadly, beyond this we really have little to no “control” over the Earth’s climate system. If the underlying natural forces that govern the bistability (or multistability) are, as they reasonably must be, slowly moving the Earth’s climate along the downhill path that will eventually force a flip back to the 80,000 year stable cold phase, destabilizing warm phase and making LIA excursions more likely and causing them to take ever longer to return to an every cooler “normal” until one finally crosses that invisible line and conditions favor positive feedback albedo-driven glacial growth back to cold phase, there isn’t a damn thing we can do about it, especially if we don’t even understand the underlying parametric dependences so we can’t even approximately predict when all of this will happen, beyond going “well, the Holocene is 11,000 or so years old, making it probable that it is going to end “soon” (within 1-2000 years) based on the recent previous interglacials and assuming that things now are like things then”.
When it does happen, they way it will happen is probably this. A Maunder Minimum will come along — we could be at the precursor of one right now, looking at a very low solar cycle max followed by a nearly flat cycle or cycles. The albedo will increase (it is already up 6%, to ballpark of 0.32 from 0.30). This will gradually drop average temperature roughly 2K. This will overwhelm any post-LIA warming CO_2 based or not, and e.g. Arctic and Antarctic ice formation will return to normal (first) and then actually increase. Glaciers that have shrunk or remained stable for the last few hundred years will grow.
If the ice/permafrost reaches far enough south/north from the poles to start to hit latitudes where it begins to materially affect the mean bond albedo, this will constitute a positive feedback effect that will “resist” glacial melting for at least some decades when the Earth’s solar-driven albedo returns to normal (one hopes) post Maunder Minimum. How long and well it resists will depend on how long the MM lasts and how much ice is laid down while it is there. The Earth will then be somewhat vulnerable, if a stable cold phase attractor has emerged beneath the current warm phase in the parameter space. If a second MM occurs (and/or we’ve stopped generating CO_2 and whatever Anthropogenic component is gradually disappearing) before the glaciers have had a chance to melt back towards normal, a second round of growth might well tip the normal sun albedo to the point that favors glacier growth.
In that event, we could easily be tipped straight over the edge to cold phase. Historical evidence is that when it happens it happens fast, and the reason is that the cold phase attractor has long since emerged as the second bistable state beneath the warm phase, so that it is just a matter of crossing that invisible line to where all natural processes conspire to push the system towards the cold phase attractor and those very forces will then do the work without any additional help. Glaciers will grow every year, increasing the albedo as they march south to further cool the system and move the glacial boundary still further south, until tropical warming balances it and further glaciation is arrested.
This would indeed be a catastrophe. We have no evidence at all that a warmer phase exists as an emergent multistable attractor. We can take a glance at the climate record of the last 50 million years (or longer) and clearly see that if anything, even the warm phase we are now experiencing is relatively unlikely (and unstable) compared to the million year norm, and living on borrowed time. If CO_2 stabilizes us in warm phase for a longer time than it might otherwise last, this is a good thing, not a bad one, because an ice age would kill 2/3 of the Earth’s population in a century between famine and war and drought unless we are able to use technology to do more with less, to learn to live in a civilized way under the conditions that existed during the last ice age.
I do predict this catastrophe, but I have no idea when it might occur. It might now be starting, and by 2112 we could be all iced up. It might not start until 3012. I do think that it is well worthwhile to do work that might illuminate our understanding to where we might be able to predict it, and/or detect its precursor signals (as I outlined above) because that would give us a small chance of actually planning for and coping with it instead of being swept along with the general kill-off of northern temperate ecology, the loss of the Canadian and Siberian and Chinese breadbaskets, the late and early killing frosts, and the inexorable march of the permafrost line south (and north) from the poles. This won’t matter to me — I’ll be dead long before the earliest dates we might gain this understanding and see the precursors — but my great great grandchildren might thank us all then for doing some good science now.
rgb
rgbatduke, do a search for Maximum Entropy Production (MEP) in non-linear / non-equilibrium fluid systems. Many papers on earth climate under MEP by Axel Kleidon, Ralph D. Lorenz and GW Paltridge, like :Entropy Production by Earth System Processes Also used in studies on other planets and regarding Ice Ages.
Interesting. As mentioned in my previous post, assuming bistability (or more), there should exist a differential description of the Earth’s stable state that is (classically) S-shaped in some parameter, with the middle branch of the S unstable (and the stability boundary between the upper and lower stable branches. The basic ODE structures that produce such a shape for open systems being externally driven are actually pretty well known (at least in quantum optics, but of course ODEs are ODEs). I would be very interested in seeing if anybody has a gross macroscopic one or two parameter model with an S-shaped stability line or surface that can at least heuristically describe glacial/interglacial transitions, especially since that line or surface would in principle predict what the stable temperature should be given any set of values for the underlying parameters, a quantity utterly missing in any GCMs or other studies and obviously one that is the sine qua non of any work that should be taken seriously. Without it one knows nothing — seriously. With it one has a foundation for understanding local fluctuations and natural variability. With that one can begin to think about looking at the effects of CO_2 and other minor drivers, given the sound evidence that a) the underlying parameter space of the gross stable branches is by far the major determinant of where we are in the climate cycle and what the temperature outside “should” be, given our non-Markovian climate history; b) natural variability around this generally stable temperature is large — order of 10-20% of the separation of the phases; CO_2 may make an important baseline contribution to the whole curve, but because it is saturated the curve itself is currently rather insensitive even to large changes in CO_2 concentration.
rgb
A widely used and well accepted but misleading terminology is “reradiate” or “re-radiate.” Typical statement such as “carbon dioxide absorbs IR then reradiates the energy,” implies that carbon dioxide does not emit IR if it does not absorb first. Radiation physics, as indicated by the SB equation, says that CO2 emits 24/7 as long as its temperature is not 0 K, regardless whether it absorbs or not.
Yeah, but quantum physics says otherwise. In this debate, quantum physics wins hands down.
If kT is less than the excitation energy of an emitting level, BB radiation will be very small indeed. This guy named Planck, you might recall, worked on this. This is why radiation from O_2 and N_2 is not a major contributor to radiative cooling of the atmosphere in spite of the fact that they are in thermal equilibrium with the CO_2 — it isn’t that they don’t have levels that can radiate, it is that those levels don’t get excited by kT-level collisions. Neither, for the most part, do those of CO_2.
If you doubt this, I refer you directly to TOA IR spectra that make the point conclusively. The CO_2 blocking of surface radiation in the IR bands, and its eventual reradiation at a much colder temperature in those bands is clearly a resonant absorption phenomenon, dominated by scattering and not simple radiative cooling. An atom has no “temperature”. A molecule has no “temperature”. It has some mix of translational and electronic energy. Radiation comes from electronic excitation energies, but most of its “thermal energy” is bound up in translation, and it is the general case that when kT is too small to populate a quantum degree of freedom, those degrees of freedom are not in thermal equilibrium with T. Again, this is all laid out in kiddie physics books, where it explains why monoatomic molecules have 3 degrees of freedom and diatomic molecules generally have 5, in spite of the fact that they should have 7 if one allows for axial rotations and vibrations. At most temperatures the axial rotation and vibratory quantum states are not excited. Tons of evidence (specific heats, etc).
rgb
What are the consensual atmospherics at Duke like? Are you a shunned outlier, or representative?
If you are referring to some sort of battle lines between skeptics and non-skeptics (concerning CAGW) I have no idea — both extremes are present, I’d guess, with the latter represented in the school of the environment but the former around as well.
I’d have to say that I am “invisible”. I don’t, after all, do active research in this (and am too busy to do more than post on WUWT although I am tempted to do a few things if/when I ever have time and support). I teach physics at the Duke Marine Lab during the summers, but have never been braced on some sort of “political correctness” issues regarding CAGW (and cannot imagine that happening at Duke, frankly) or threatened with non-employment in what is an elective gig. Of course I am a damn good teacher of physics (if I do say so myself:-) and both the ML admins and my summer students are extremely pleased with the course I teach there.
I don’t hide my skepticism from students, but I’m not teaching “climate physics” and I don’t generally spend hours going over it all with them, either. I simply direct them towards a few true facts and point out that the “consensus among scientists” isn’t, and that physicists are in any even highly consensus proof, given that our entire discipline has been found to be completely incorrect on multiple occasions in the last 400 years, forcing paradigm shifts to completely new conceptual foundations. Physicists don’t believe in “settled science”, they believe in science that appears to be pretty reliable, so far. It’s better that way. Especially when making egregious statements about the most difficult coupled Navier-Stokes problem in the world in a context where we cannot even predict is gross zeroth order normative behavior or identify all of the critical parameters.
rgb