This page is a blog article in progress, written by Tim van Beek.
When we talked about putting the Earth in a box, we saw that there is a gap of about 33 kelvin between the temperature of a black body in Earth’s orbit with an albedo of 0.3, and the estimated average surface temperature on Earth. An effect that explains this gap would need to
1) have a steady and continuous influence over thousands of years,
2) have a global impact,
3) be rather strong, because heating the planet Earth by 33 kelvin on the average needs a lot of energy.
Last time, in a quantum of warmth, we refined our zero dimensional energy balance model that treats the Earth as an ideal black body, and separated the system into a black body surface and a box containing the atmosphere.
With the help of quantum mechanics we saw that:
Earth emits mainly far infrared radiation, while the radiation from the sun is mostly in the near infrared, visible and ultraviolett range.
I claimed that only very special components of the atmosphere react to infrared radiation. Not the main components $O_2$ and $N_2$, but minor components with more than two atoms in a molecule, like $H_2 O$, $O_3$ and $CO_2$. These gases absorb and re-emit a part of Earth’s emission back to the surface. Today I would like to expain the reasons for this in a little bit more detail than last time.
The downward longwave radiation (DLR) emitted by infrared active gases leads to an increased incoming energy flux from the viewpoint of the surface.
This is an effect that certainly matches the points 1 and 2 above: It is both continuous and global. But how strong is it? What do we need to know in order to calculate it? And is it measurable?
There has been a lively - sometimes hostile - debate about the “greenhouse effect” which is the popular name for the increase of incoming energy flux caused by infrared active atmospheric components, so maybe you think that the heading above refers to that.
But I have a different point in mind: Maybe you heard about guiding systems for missiles that chase “heat”? Do not worry if you have not. Knowlegeable people working for the armed forces of the USA know about this, and know that an important aspect of the design of aircrafts is to reduce infrared emission. Let’s see what they wrote about this back in 1982:
The engine hot metal and airframe surface emissions exhibit spectral IR continuum characteristics which are dependent on the temperature and emissivity-area of the radiating surface. These IR sources radiate in a relatively broad wavelength interval with a spectral shape in accordance with Planck’s Law (i.e., with a blackbody spectral shape). The surface- reflected IR radiation will also appear as a continuum based on the equivalent blackbody temperature of the incident radiation (e.g., the sun has a spectral shape characteristic of a 5527°C blackbody). Both the direct (specular) as well as the diffuse (Lambertian) reflected IR radiation components, which are a function of the surface texture and the relative orientation of the surface to the source, must be included. The remaining IR source, engine plume emission, is a composite primarily of $C0_2$ and $H_20$ molecular emission spectra. The spectral strength and linewidth of these emissions are dependent on the temperature and concentration of the hot gaseous species in the plume which are a function of the aircraft altitude, flight speed, and power setting.
This is an excerpt from page 15 of
You may notice that the authors point out the difference of a continuous black body radiation and the molecular emission spectra of $CO_2$ and $H_2 O$. The reason for this, as mentioned last time in a quantum of warmth, is that according to quantum mechanics molecules can emit and absorb radiation at specific energies, i.e. wavelengths, only. For this reason it is possible to distinguish far infrared radiation that is emitted by the surface of the Earth (more or less continuous spectrum) from the radiation that is emitted by the atmosphere (more or less discrete spectrum).
Last time I told you that only certain molecules like $CO_2$ and $H_2O$ are infrared active. The authors seem to agree with me. But why is that? Since last time we had some discussions about whether there is a simple explanation for this, I would like to try to provide one. When we try to understand the interaction of atoms and molecules with light, the most important concept that we need to understand it that of an electric dipole moment
Let us switch for a moment to classical electrostatic theory. If you place a negative electric point charge at the origin of our coordinate system and a positive point charge at the point $\vec{x}$, I can tell you the electric dipole moment is a vector $\vec{p}$ and that:
For a more general situation, let us assume that there is a charge density $\rho$ contained in some sphere $S$ around the origin. Then I can tell you that the electric dipole moment $\vec{p}$ is again a vector that can be calculated via
So that is the definition of how to calculate it, but what is its significance? Imagine that we would like to know how a test charge flying by the sphere $S$ is influenced by the charge density $\rho$ in $S$. If we assume that $\rho$ is constant in time, then all we need to calculate is the electric potential $\Phi$. In spherical coordinates and far from the sphere $S$, this potential will fall of like $1/r$ or faster, so we may assume that there is a series expansion of the form
When our test charge is far away from the sphere $S$, only the first few terms in this expansion will be important to it.
In order to completely fix the series expansion, we need to choose an orthonormal basis for the coordinates $\phi$ and $\theta$, that is an orthonormal basis of functions on the sphere. If we choose the spherical harmonics $Y_{l m}(\phi, \theta)$, with proper normalization we get what is called the multipole expansion of the electric potential:
$\epsilon_0$ is the electric constant.
The $l = 0$ term is called the monopole term. It is proportional to the electric charge $q$ contained in the sphere $S$. So the first term in the expansion tells us if a charge flying by $S$ will feel an overall net attractive or repulsive force, due to the presence of a net electric charge inside $S$.
The terms for $l = 1$ form the vector $\vec{p}$, the dipole moment. The next terms in the series $Q_ij$ form the quadrupol tensor. So, for the expansion of the potential we get
For atoms and molecules the net charge $q$ is zero, so the next relevant term in the series expansion of their electric potential is the dipole moment. This is the reason why it is important to know if an atom or molecule has states with a nonzero dipole moment: Because this fact will in a certain sense dominate the interactions with other electromagnetic phenomena like light.
If you are interested in more information about multipole expansions in classical electrodynamics, you can find all sort of information in this classical textbook:
In quantum mechanics the position coordinate $\vec{x}$ is promoted to the position operator; as a consequence the dipole moment is promoted to an operator, too.
A rough estimate of energy levels for molecules shows that
electron energy levels correspond to ultraviolett and visible light,
vibration corresponds to infrared light and
rotation corresponds to microwaves.
Tim van Beek: Heuristic explanation for the different energy levels?
For atoms and molecules interacting with light, there are certain selection rules. A strict selection rule in quantum mechanics rules out certain state transitions that would violate a conservation law. But for atoms and molecules there are also heuristic selection rules that rule out state transitions that are far less likely than others. For state transitions induced by the interaction with light, a heuristic transition rule is
Transitions need to change the dipole moment by one.
This selection rule is heuristic: For one, it is valid when the radiation wavelength is bigger than the molecule, which is already true for visible light.
Secondary, transitions that change the dipole moment are far more likely than transitions that change the electric quadrupole moment only, for example. But: If an atom or molecule does not have any dipole transitions, then you will maybe still see spectral lines corresponding to quadrupol transitions. But they will be very weak.
Molecules that are infrared active need to have vibrational modes that have a nonzero dipole moment. But if you look close enough you will find that molecules that are not “greenhouse gases” can indeed emit infrared radiation, but the amount of radiation is insignificant compared to that of the greenhouse gases.
If you take a look at molecules consisting of two atoms of the same species like $O_2$ and $N_2$, you will find that such molecules can never have any vibrational states with a dipole moment at all, which means that more than 99% of all molecules in the atmosphere have an insignificant contribution to infrared radiation.
Tim van Beek: Compare black body radiation to the emission spectrum of CO2 and H2O.
If you would like to learn more about molecules, have a look at:
If you speak German and are interested in a very thorough and up to date treatment you could try:
and
So, if we try to calculate the DLR effect for the atmosphere of the Earth, it is sufficient to focus on molecules with vibrational modes with a nonzero dipole moment.
The next important point is mentioned by the air force authors in this statement:
The spectral strength and linewidth of these emissions are dependent on the temperature and concentration of the hot gaseous species...
Of course the temperature, pressure and concentration of atmospheric components are not constant throughout the whole atmosphere, so this is a point that is also important for us when we investigate the radiation properties of the atmosphere.
But why is this important? Temperature and pressure change molecular emission spectra by line broadening mechanisms. The concentration of gases is important because saturation effects lead to a non linear dependence of absorption and emission on the concentration.
The most important mechanisms of line broadening are:
Tim van Beek: List of line broadening mechanisms.
We will need to calculate thermodynamic properties of the atmosphere, at least approximately, to determine the molecular emission spectra.
Tim van Beek: Adiabatic lapse rate and why this does not explain the 33 kelvin gap.
What about the dependency of emission and absorption on the concentration of infrared active gases?
Tim van Beek: Explanation of the log dependency.
Tim van Beek: Equations of radiation transfer of the atmosphere (at least a simple approximate version of it). Why it is too complicated to solve by hand.
Calculating the DLR on a sheet of paper, even with the use of a pocket calculator, would quickly turn out to be quite a task.
We could start by making some assumptions about the different layers of the atmosphere. We would also need to look up molecular spectra. Thankfully, for this task there is some help: The HITRAN database. HITRAN was founded by the US air force. Why? I don’t know, but I guess that they needed the data for air craft design, for example. Look out for the interview with Dr. Laurence Rothman for some background information about HITRAN; there is a link to it on the home page.
But anyway: We see that this task is complicated enough to justify the effort to write computer code to handle it. But we are lucky: Not only have others done this already for us. In fact you can find a survey of some of the existing software programs on Wikipedia:
A kind soul has provided a web interface to one of the most prominent software programs, MODTRAN, for us to play around with:
Now that we have some confidence in the theory behind DLR, we are ready to look into measurements.
To measure DLR and check that it is really the energy flux coming from infrared active components of the atmosphere and not some strange artifact, we have to
point some measurement device to the sky, to measure what goes down, not what goes up and
check that the spectrum we measure consists of the characteristic molecular spectra of $CO_2$, $H_20$ etc.
The kind of measurement device we could use for this is called pyrgeometer, for pyr = fire and geo = earth.
For starters we should look for conditions where there is minimum radiation from other sources, no clouds and only a small amount of water vapor. What would be a good place and time on Earth to go to? A dedicated team of scientists decided to weather the grim conditions of the antarctic during polar night for this purpose:
Also:
Also:
Tim van Beek: I would like to add radiation measurements, maybe some can be found here:
Als
Just to have a number, the flux of DLR (downwards longwave radiation) is about 300 $W m^{-2}$.