Need Help: Can You Model CO2 as a Greenhouse Gas (Or is This Just Wishful Thinking?)

Andre

Right, but it still requires more density for air to descent, if that air is containing more energy it must at an higher ambient temperature and/or a higher pressure. If it is at an higher pressure, it will expand and decreases in density increasing it's bouyancy, stopping the downdraft. I still can't see how in a complete convection cell the net energy flow can be downwards.

There is still more air convecting up also transfers (heat) energy into potential energy, which process is reversed in descending air

Not if one is a declared crook according to the moral panic principle.

Here is the full quote

A subsidence inversion is the norm above the deserts, in the downdraft regions of the hadley cells, as the descending air increases in density and heats up adiabatically, additional convection would be extremely rare and certainly not the norm.

vanesch

Thanks for the explanations. I will try to digest them "out loud".

For all clearness, I'm not talking about earth's atmosphere (yet), only about "plateworld"s atmosphere: a black hot plate, a gaslayer on top of it, and outer space. Some magic to do things with the gas, which is normally not possible, such as switching on and off convection, radiation, and a few other things, to get an understanding of the different mechanisms and their interplay.

We start by giving our gas layer (with some magic, or gravity) a pressure profile, with pressure decreasing with altitude. I will try to see where I get.

Ok, so what's understandable from this, is that if the pressure profile is given, the relative temperature curve is given if there is sufficient convection. A balloon with some gas at temperature T1, lifted in this atmosphere, will cool down adiabatically (because the pressure lowers, and the balloon expands). A balloon going down will heat up (compression). If the atmosphere is not heated or cooled (no radiative stuff in it), it would reach a certain equilibrium given by this adiabatic.

Let's play a bit with this non-radiative atmosphere. Let us say that at the surface, I've 10 degrees, and at 20 km, I have -50. (making numbers up here).
Now, suppose that with electric heaters, I bring the layer at 20 km at -30. This would mean that the less cold air at 20 km, going down in the convective stream, will now bring the surface layers to a much higher temperature (say, 40 degrees, following the adiabat from -30 and compressing). The whole atmosphere will now settle to a new equilibrium, again with an adiabat, but with the top layer now at -30, and the surface at 40.

Right. This is something I didn't realise that convection could transport heat down against a temperature gradient.

Let's play another game (this is fun!). Suppose that thermal conductivity of our atmosphere is very bad but not 0. Still no radiative stuff, we're just looking at the transparant atmosphere. We switch off the EM field (I told you we had magic!).

Now, we do the following: our initial surface is at 10 degrees, the top of the atmosphere is at -50, and there is this adiabatic equilibrium due to convection (which is driven also by magic).

Suppose now that we build a huge heat exchanger at 20 km height, and another at the surface. Suppose that the surface has a thermostat that keeps it at 10 degrees, but heat can be supplied or extracted. It's a thermal reservoir. Now, we connect our two heat exchangers with some or other liquid. We take heat from the soil at 10 degrees, and bring it to the upper layers to heat the upper layers. This is possible, because up there, it is -50.
We do this until the upper layer is now at -30. We are in the same situation as before, so now the lower part of the atmosphere is hotter than the surface !

There is something wrong. We violated the second law here: we took heat from the surface at 10 degrees, delivered it to our gas at -50 (still ok), and this heated the air to 30 degrees just above the surface. So the whole cycle took heat at 10 degrees and delivered it at 30. That's against the second law (unless we do work). So, the problem was that I introduced too much magic, and introduced convection even when the upper temperatures were above the adiabat. That forced convection (my magic) did work on the gas.

I guess that if upper layers, in one way or another, are hotter than they should be according to the adiabat, convection simply stops.

So it seems that you can't heat "downwards" using convection, no ? Violates the second law, no ?

So where's the culpritt ? I would guess that it comes from thinking that a hotter gas can convect down in a cooler gas. It will be less dense, so it will have tendency to go up, not down.

So, convection cannot really take heat down, can it ? In other words, the adiabat is defined by the temperatures in the lower layers, not in the upper layers. Am I right here ?

So, is this true ? What about my above example ?

Or is it rather: convection will transport heat up, and if it should transport heat down, it stops.

I will stop here already (didn't know it when I started typing) because I'd rather sort this out clearly before going on.

edit: I was typing this independently from the discussion with Andre, but it seems he butted on a similar difficulty after reading the exchange...

Evo

He said impartial Ivan.

chriscolose

Vanesch,

What's happening with the greenhouse and convection can be understood at a wide variety of levels. In actuality, constructs like S(1-a)/4 = sig T^4 are not actually used in sophisticated GCM's, but such simple formulas valuable way of explaining the basic physics of radiation balance, and serve as a bridge between grey-gas models and more realistic models.

One classical "layer model" which is often employed at a lower level allows one to think of several imaginary "panes of glass" floating in the atmosphere which are perfectly transparent to visible and perfectly opaque to infrared radiation.



One can then proceed to set up multiple equations and go about solving for the temperature at each layer. Generalizing, the surface temperature will end up being the top temperature (the emission layer) multiplied by (n + 1)^0.25 where n is the number of layers. So, a two layer atmosphere will have a surface temperature of 335 K. This suggests that radiative equilibrium is not a good approximation for the surface temperature, which loses substantial heat by convection and conduction as well. With radiative equilibrium, the lapse rate of temperature too large in the troposphere, the stratosphere is approximated pretty good, but the surface is too hot. With other forms of heat transfer now-- The whole troposphere is well mixed in heat, and is more or less constrained by convection to stay near the moist adiabat. In that sense, the vertical structure is largely fixed by convection and the IR heating simply sets the intercept (e.g. the lower tropospheric temperature).


In actuality, the atmosphere is semi-transparent to a differing degree at different wavelengths. The radiative transfer issue is best addressed numerically with sufficient number of vertical layers to resolve the atmospheric temperature and absorber distributions, and with sufficient resolution to pin down the spectral dependence of individual gases. Looking down from space you would indeed see radiation coming from various levels of the atmosphere, but the bulk of it comes from some location (determined by the atmospheric greenhouse composition) where opacity is strong. This is often called the [tex]\tau = 1[/tex] level and much of the radiation below here is absorbed before getting to space and much radiation from above is a small term as emissivity is weak. The effect of adding CO2 is to raise [tex]\tau = 1[/tex] to higher altitude (lower pressure) thereby warming the whole troposphere. So while a real "effective layer" doesn't exist, it's a usefuil concept for thinking about the radiation balance of the planet.

vanesch

Yes. However, if you take "black sheets" each time, with no radiation that transmits through a layer without being "thermalized", then you have actually a series of independent "resistors" (except that "ohm's law" is not linear but for small temperature diffs we can linearise).

If we take one such element, with on one side T1 and on the other, T2, we have a net power transmission between them of sigma (T1^4 - T2^4).

Assuming small temperature differences, we have approximately 4 sigma T^3 (T1 - T2), and we can roughly say that one such "interlayer" corresponds to a thermal resistance of
[tex] \frac{1}{4 \sigma T^3} [/tex]

(current = 1/R x potential difference)

The different successive layers are series connections of these resistors.

So the more of these the radiation has to cross, the higher the total resistance, and hence the higher the temperature difference for the same thermal flux ("current").

The more radiatively absorbing gasses you have, the more of these "black" layers we have.

In fact, you also have to count the last "gap" towards outer space as a resistor of the kind, but here you can for sure not linearize anymore as T2 = 0 (or 4 K if you want to).

So this explains intuitively the greenhouse effect in a layered, static atmosphere.

I'm trying to wrap my mind around what is the influence of convection in this picture.

sylas

(In what follows, my main reference is an online undergraduate level textbook, Principles of Planetary Climate, by Ray Pierrehumbert. You can find the same material in other texts; this one has the advantage of being easily available as a shared reference by anyone who wants to look it up. If I refer to page numbers or equation numbers; they are from this text. I take full responsibility for any errors in my answers, and you can check the methods I apply with this reference.)

I have made an error in my posts; I spoke of convection working to maintain the lapse rate. That is incorrect; I should have said convection works to decrease a lapse rate towards the adiabat, or near it. A weaker or negative lapse rate is stable against convection, so convection in the atmosphere only heads towards this point from one direction.

We've been considering the case of an "optically thin" atmosphere, with minimal thermal emissivity. If there is no energy exchange with the surface, then the atmosphere in this simple case would be isothermal (one temperature) at 2^-0.25 = 0.84 of the surface temperature. (p 142) This is called the "skin temperature". But because there is an energy exchange where the atmosphere is in contact with the surface, the bottom of the atmosphere is heated by the surface; and this proceeds up the atmospheric column to establish a temperature gradient, up to the point where the "skin temperature" is again established, and from there you get an isothermal stable stratosphere (p 143). Radiant energy transfers make life more complex; but the radiatively inert case in this example is simpler. Here's a diagram from the book.


The lapse rate is therefore directly linked to the height of the tropopause, given a surface temperature balanced with solar input, and a stratosphere at the skin temperature. If the tropopause is at low altitude, the mean lapse rate is large; and unstable. This leads to heating, by transfer from the surface and then by convection; that raises the tropopause, until you get to the adiabatic lapse rate, which is now stable.

Given the small loss of energy from the atmosphere, on the assumption of being optically thin, or radiatively inert, this equilibrium state has negligible net flow of energy up or down, and that does mean convection processes will sometimes transport energy up, and sometimes down, however this is presuming that you have a bit variance in conditions, rather than always being right at an equilibrium.

I think you really only get energy flux downwards by "forced convection", or a mechanical result of wind or other movements that do work. But I'll accept guidance from others on this point. The major point is that the net upwards energy flux into a radiatively inert atmosphere is zero.

Net isn't downwards. In the radiatively inert case, the net is effectively zero. But that suggests that as you introduce a bit more complexity, like horizontal wind and so on, there's going to transient periods of a downward flux of energy, occasionally, with other periods of a net upwards flux.... sometimes up, sometimes down. I know that a downwards energy flux is not a stable situation.

And in particular, any implication I gave that you get a spontaneous movement of energy downwards against the temperature gradient of the conventional lapse rate was my mistake. Such movement may occur, I believe, as a result of mechanical work from winds, but they are not sustained.
I don't panic, and will be happy to back you up if you get unfairly disciplined for simply being presumed to be in error. I do appreciate your substantive engagement, whether I agree with it or not.

Um... with respect, you continue to omit the initial sentences of the paragraph which describe the lapse rate being assumed. I have taken the liberty of inserting the rest of the paragraph in bold, where the lapse rate is mentioned explicitly; and adding the link to the post through the quote tag.

It's quite true that you can get inversions within the troposphere. They tend to be of a limited depth; less than a kilometer. The majority of the troposphere in a real planet is still with the positive lapse rate (falling temperature with altitude) and this is the case in the simple example I was explicitly discussing.

The point is that in the troposphere, any layer tends to be gaining radiant heat with respect to lower levels, and losing it with respect to higher levels, by virtue of the temperature gradient that occurs in the troposphere, and so you can't presume net heating or cooling immediately.

However, I can't quibble too much here, because the big error here in my post is in the second paragraph you've quoted. Convection does not necessarily work to oppose heating or cooling. It is only an overly large lapse rate above the adiabatic rate that is unstable to spontaneous convection.

This doesn't alter the main point that increasing the capacity of an atmosphere to interact with thermal radiation will give higher temperatures at the surface and in the troposphere; even though the normal equilibrium at those higher temperatures may show radiative cooling at that level, balanced by the special heat flux.

But it's still another screwup, and I am glad to acknowledge it and fix it. Thanks.

A nod to vanesch as well for picking up the problem also; you get the kudos for being first, and I sneak in third.

Cheers -- sylas

Andre

You're too kind Sylas

Would the band saturation decrease that effect? If that frequency band is 'saturated' it appears that it won't make that much difference anymore how many times radiation energy is absorpted and re-emitted.

Ah let's try some ideas, especially with wet convection, involving latent heat. So, as a wet surface heats up, water evaporates (latent energy -which reduces the temperature increase). Conduction and radiation heat up the lower layer(s) of the troposhere, causing the well discussed convection. Heat- and latent energy -water vapor- are now transported up. Due to expansion the updraft cools adiabatically and water condenses forming clouds and releasing the latent heat again. Clouds are good radiators as they radiate on all water IR- frequencies. So this energy is radiated outwards in al directions as it would have done on the Earth surface without convection. But the difference is that energy -on water frequencies) emitted upwards will find less water vapor molecules because the upper levels are much drier than the surface levels. Evidently, the CO2 frequency bands are also less relevant here. Consequently the energy emitted by clouds (tops), on water frequencies, has more chance to escape into space than energy emitted by the surface in all bands including the CO2 frequencies.

Now if the greenhouse gas concentration was to increase then the heating of the lower atmosphere by radiation was also to be increased, this would enhance the convection rate, transporting more energy upwards, where more energy can radiate into space. Consequently it appears that convection acts as a negative feedback on greenhouse gas variation

Skyhunter

This makes no sense. Absorption and re-emission makes a difference everytime it happens.



Why do you say it is evident that CO₂ absorption is less relevant at higher altitudes?

If the emission frequency is in the active CO₂ bands, (which you are saying it is) then it will be absorbed, since CO₂ is well mixed.

I agree that the scarcity of WV in the upper troposphere and stratosphere leaves larger windows for radiation to escape into space. But the concentrations of CO₂ are fairly uniform.

sylas

(As before, any references are pages or equations in Principles of Planetary Climate)

Actually, it does make a difference; because of the lapse rate.

The example described here is completely saturated, even with a single "pane"; the stated assumption is that each successive pane is completely opaque to the upwards thermal radiation. Each pane is warmer than the one above it, and the more successive panes you have, the higher the temperature of the bottom pane; since the uppermost is the one that it at the effective radiating temperature to balance to short wave input.

This kind of effect is seen, for example, in the atmosphere of Venus, which is profoundly saturated. The topmost level of the atmosphere on Venus is at the effective radiating temperature... which is actually colder than Earth, because Venus has a very high albedo. Despite being closer to the Sun, Venus actually absorbs less solar energy per unit area than Earth! It is so hot because of a super greenhouse effect; thermal radiation is absorbed and re-emitted many times up that dense carbon dioxide atmosphere, and all the way the lapse rate is maintained, so that right at the bottom you are far hotter than the effective radiating temperature at the top of the atmosphere.

The big omission of this example is convection, and vanesch also asked about that. In a profoundly optically thick atmosphere like this, there is a natural radiative lapse rate, which corresponds directly to the successively lower temperature on higher panes of glass in our example. At the same time, there is also the natural convective lapse rate, which is determined by the adiabat. Convection will be at work if the radiative lapse rate is greater than the adiabatic lapse rate. In that case, convection will relax the lapse rate, and that will reduce the temperature difference between top and bottom from the purely radiative case. I'll consider than some more later on.

In the meantime, note that the Earth is rather different to Venus. (p253) On Venus, increasing greenhouse gas concentrations works mainly by raising the emission altitude. On Earth, increasing concentrations works mainly by widening the saturated bands; additional absorption occurs in the "wings" of those bands, more than by raising the emission altitude of the saturated regions.

Cheers -- sylas

BrianG

The greenhouse effect is so weak, it's impossible to duplicate AGW in the lab or the field. These experiments aren't peer reviewed, can't be duplicated and no responsible researcher or lab has claimed credit.

sylas

The original experiments which demonstrated how the greenhouse effect works with directly measured temperature differences were conducted about 150 years ago by John Tyndall, one of the great Victorian experimental scientists. His experiments are described in message msg #10 of this thread, with a link to John Tyndall's book online that describes them in more detail.

In fact, the greenhouse effect is very strong indeed, and it is responsible for the Earth having a livable climate at all. This is not in any credible dispute, and is widely discussed in basic text books dealing with the Earth's climate. Extracts from John Tydall's lecture on this are presented in msg #76.

The nineteenth century was a productive period in experimental physical science; and the basics of thermodynamics and temperature were established then, and have been extensively developed since. Thermodynamics at this level is not in the slightest physical doubt, and continues to be given in elementary text books on the subject.

In the modern era, you are not likely to find much in the way of experimental work specifically measuring temperature change, except in high schools or undergraduate lab work. A number of such experiments have been given in the thread, and they can easy show a temperature difference in controlled conditions. A selection of such experiments is given in msg #59.

The actual strength of the greenhouse effect on Earth is well known from basic comparison of the radiation measured from the surface, and from what escapes to space. The total effect is about 33 degrees Celsius (about 60 degrees Fahrenheit). The calculation of this magnitude is shown in msg #96. The actual amount of radiation coming to the Earth from the atmosphere is very large... hundreds of watts per square meter, day and night. This is measured directly, and has been for 50 years. Citations for such measurements are given in msg #64.

Brian has not actually directly addressed any of these experiments or calculations or measurements. He has, however, insisted that an experiment should use concentrations of carbon dioxide equivalent to that in the atmosphere... which is a bit less than 1/2500 by volume. But he wants to see the effect in a lab... in a few meters of air. This is, of course, absurd; and that has been pointed out. He's effectively demanding to see the greenhouse effect using much less than one thousandth of the actual amounts of greenhouse gases that apply to give us a livable climate. This is explained in msg #85.

There's nothing wrong with disagreement over fundamental points. We can explain the relevant physics. But at this point, Brian has long since stopped engaging the discussion and the evidence, and has taken to repetitive posting of a couple of lines that just make the same point which has been demolished many times over in the thread.

For the record -- sylas

BrianG

I directly address the ESPERE experiment you cite:

http://www.espere.de/Unitedkingdom/w...greenhouse.htm

The data is extremely limited, reduced to one twenty minute run with four data points for each of the two samples. The work isn't peer reviewed, isn't attributed to any specific lab or principle researcher.

sylas

Of course it isn't. It's a simple experiment intended for students; not a research report. This has been explained many times now. I don't think you'll find such basic teaching experiments in a formal scientific paper any more. It's not that trivial to get a scientific paper published, you know!

Experiments to measure temperature directly from thermal interaction of a gas with radiation have not been particularly important now for over a hundred years. Those experiments can be repeated -- and ARE repeated, as shown in these simple teaching experiments -- and they do show temperature effects very easily.

The state of science now is that the basic thermodynamics is nailed down solidly, and what is important for physics is measuring the properties of how light and matter interacts; how energy is absorbed and emitted. Even the spectral emissivity of CO2 -- which is the quantity of relevance -- is not something determined by experiment any more.

But for some reason you have dismissed those details, saying you don't dispute spectral characteristics. Apparently, you just dispute the basic consequences which follow from this when applied to something too big to fit in a lab.

Trying to reproduce the entire atmosphere is not something you do in a lab. Trying to have a couple of meters of gas with 500ppm CO2 has less than a thousandth of the effect of an atmosphere -- yet that is apparently what you think should be done. It's not a sensible experiment. It bears little relation to an entire atmosphere, and it doesn't just scale simply even if you could measure the tiny impact as a temperature. You can EASILY get a temperature difference using an amount of CO2 similar to that in the atmosphere -- but that doesn't scale easily either, because of the importance of lapse rate that has been discussed.

You most certainly can measure the energy effects of even small amounts of CO2; but you don't do that by measuring temperature. You measure the radiation directly.

For some reason which you have not explained, you apparently don't think that is good enough.

Here's another example of what a relevant experiment does in modern physics. (And note that even THIS is simply a confirmation of the basic quantum theory used to calculate the interactions of light and matter.)

• Roy, S. et al. (2002) Mid-infrared polarization spectroscopy of carbon dioxide, in Applied Physics B: Lasers and Optics, Volume 75, Number 8 / December, 2002.

Here are the basic facts.

• The experimental measurement of the effects of gasses on thermal radiation, determined by measuring temperature effects directly, were conducted by John Tyndall in the mid nineteenth century. Nobody, ever, has claimed that these experiments are somehow incorrect or don't measure what is described. That would be absurd.
• Similar experiments continue to be conducted now, although in modern days they are teaching experiments used in schools. It's not something you bother with in a research paper.
• To measure temperature effects, you need a fair amount of carbon dioxide to absorb sufficient thermal radiation to have an impact. That means you can't do it usefully with 500ppm CO2 in a lab. You can measure the backradiation from the sky directly. It is very large, and has been cited in the thread. Or you can do experiments like those that have been described.
• The impact of CO2 on energy transmission continues to be studied. You measure the radiant energy directly, as this gives you much better resolution than trying to measure temperature. That some individuals are apparently dubious of how energy and temperature are related is not the problem of working scientists, but is rather a problem of education. That is what I am trying to help with here.

Sylas

Andre

Interesting and straight forward, but isn't half the story? What happenes if the light is turned off? After all, the sun shines only half a day.

BrianG

You say, "Trying to reproduce the entire atmosphere is not something you do in a lab." then you cite an experiment that would necessarily be faulty:

http://www.espere.de/Unitedkingdom/w...greenhouse.htm

And you say this is a teaching tool in schools? What are they teaching, bad science?

Then you insist we ignore temperature differences and look at spectroscopy.

The preview to your link says nothing about the results of varying concentrations of CO2.

http://www.springerlink.com/content/phcvdcmce4y2hff7/


Our key question is, what does a few parts per million of manmade CO2 emissions do to atmospheric temperature? Not, what color is the sky.

sylas

The combination of CO2 and H2O as the two major greenhouse gasses results in about 33 degrees of additional warmth over what you would have otherwise.

It's not possible to divide the warming between the two gases, or other smaller contributors, as if in a linear sum. Each one in part compensates for the other; so that the impact of the two together is less that the sum of each one acting alone. But as a rough comparison it is fair enough to say that H2O is roughly twice as significant as CO2 for the total greenhouse effect on Earth.

The 33 degree impact of Earth's greenhouse effect is a comparison of what temperature we actually have and what temperature would be required to radiate directly into space, as occurs on the Moon. This was explained earlier in the thread. This also fits with the hundreds of watts per square meter which is measured as radiation coming to the surface from the atmosphere; energy that would not be available without gases in the atmosphere to radiate it. The spectrum of this backradiation aligns with the bands where Earth's greenhouse gases are active; and measurement of the emission spectrum from above the atmosphere also shows clearly the major bands where greenhouse absorption occurs, and the emission is coming from the higher cooler altitudes in the atmosphere.

Andre asks what happens at night. The answer is that the greenhouse effect has a hugely significant role for keeping nighttime on Earth a moderate temperature. The atmosphere has a substantial capacity to hold energy as heat, and so it keeps emitting thermal radiation all night, and this is a major source of warmth. This is measured directly in the experiments I cited previously of backradiation, which includes night and day measurements.

----

One problem with this discussion is that it can easily be mixed up with the idea of global warming. In fact, this is quite a distinct problem.

Global warming and climate change is not simply about the greenhouse effect -- it is about the impact of CHANGES to composition of the atmosphere. What happens when you get additional greenhouse gases in the atmosphere? There are a host of real and interesting scientific questions related to this -- and all of that debate is sidestepped with what is essentially an irrelevant distraction in this rejection of the very idea of any greenhouse effect at all.

This is why I said earlier that denial of the greenhouse effect is comparable to creationism. I did not mean that as an attack on individuals, but as a characterization of the scientific argument itself. All the ideas about the genuinely open questions -- like the measure of climate sensitivity, or the impact of other non-greenhouse forcings, or the regional distribution of impacts from a changing climate and constraining of feedbacks and much else besides is a whole different level.

Suggesting that the the greenhouse effect doesn't exist, or that CO2 or H2O have no effect on temperature by virtue of thermal emissivity, or that there's some source of energy somehow giving the surface the extra 33 degrees over Earth's effective emission temperature into space other than the measured heat from the atmosphere, is all really an attack on fundamental thermodynamics established in the nineteenth century and now quite fundamental in physics education; and absolutely misses completely any serious examination of the genuinely open questions in climate.

Cheers -- sylas

skypunter

What happens when water vapor is carried aloft and is condensed?
Isn't a great deal of heat radiated above the densest layers of greenhouse gases?
Wouldn't this result in a large net outflow of energy?
How is this seemingly random thermodynamic process handled by computer models?