That honestly doesn't make any sense to me. How could heat pass through a gas one way but not the other? Its not like our upper atmostphere has a bunch of doors that can only open one way. To me, that seems to fly in the face of everything I've been taught about how the physical universe works.

If heat can pass through a substance one way, there's no reason why it can't go through the other way. Except for 'greenhouse gases' apparently.

The closest thing I know of that works this way is one-way mirrors. But that doesn't have the same effect. The reason you can see through one side and not the other is because most light is reflected, but not all. Thus if you're in a dark room, the light coming in from the outside will overwhelm what little light is being reflected from within the room. But if you're on the outside, virtually no light is coming through from the dark side and thus you can only see what is being reflected.

In short, it makes no sense to me and flies in the face of everything I've been taught. Now, I'm not a climate change denier. Actually, this is something I just now thought about. I've never really bothered to question it before.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – gerrit
    Commented Jul 30, 2018 at 14:00

3 Answers 3


In a nutshell:

The radiation that enters is shortwave radiation from the sun. Solar radiation is dominated by visible (as well as UV and near infrared) radiation with a wavelength mostly between 0.2 µm and 2 µm. This wavelength is determined by the temperature of the Sun, in the order of 6000 K. For visible radiation (roughly between 0.4 µm and 0.7 µm), all the gases in the atmosphere are (almost completely) transparent. Clouds do scatter this radiation so you can't see the Sun directly when it's cloudy, but they still let plenty of light through, so it's still light on a cloudy day (just less bright), although it does feel cooler.

The radiation that leaves is longwave radiation, also known as terrestrial radiation or (incorrectly) as thermal radiation. For Earth, this is dominated by radiation with a wavelength of mostly between 4 µm and 40 µm. This wavelength is determined by the temperature of the Earth, at around 290 K. At these wavelengths, greenhouse gases are mostly opaque. They absorb the radiation. The greenhouse gases reradiate the absorbed heat, but in both directions (back to Earth AND up to higher layers in the atmosphere), and at a lower temperature. The upward longwave radiation gets in turn absorbed by higher-up layers of greenhouse gases, et cetera, such that ultimately the layers of greenhouse gases that do radiate into space tend to be high up in the air (say at 240 K). Thus, their overall effect is to radiate less heat into space than the Earth surface would in the absence of those gases.

The most important greenhouse gases are water vapour, carbon dioxide, and methane.

Atmospheric opacity
Source: gisgeography.com

Planck curves
Source: science of doom

  • 8
    $\begingroup$ @Barry Carter: It's not really a matter of reflection. Reflected photons would (simplistically) go right back out (which, after all, is why you can see Earth from space). It's the absorbed photons that matter: they heat the material that absorbs them. Any heated material emits infrared photons. $\endgroup$
    – jamesqf
    Commented Jul 28, 2018 at 17:54
  • 3
    $\begingroup$ @Barry Carter: Yes, the energy that comes in as visible light photons gets absorbed, heating up whatever absorbs it. And any warm object emits IR photons, cooling itself in the process. The IR energy is emitted all the time, it's just that during the day there's more coming in. So for instance a rock in the sun heats up during the day, then cools off at night. $\endgroup$
    – jamesqf
    Commented Jul 28, 2018 at 19:52
  • 6
    $\begingroup$ -1, until you fix the incorrect sentence (and the sentences that follow) that starts with "Since the greenhouse gases are mostly high up in the air ..." Except for ozone, the greenhouse gases, like most of the atmosphere, is mostly low in the air. And it's not ozone that is the problem with regard to global warming. It's mostly carbon dioxide and methane, and their effect on water. CO2 and methane are well-mixed (i.e., mostly in the lower atmosphere) and water is almost exclusively in the lower atmosphere. $\endgroup$ Commented Jul 28, 2018 at 22:59
  • 2
    $\begingroup$ Excellent answer. It might help the OP (like it helped me) to think about how objects change colour when they get hotter. Metal glows red, then yellow, then becomes "white hot" as its temperature increases and the emitted wavelength changes. On a gas burner, the hottest part of the flame is blue (or even invisible/UV). $\endgroup$
    – craq
    Commented Jul 29, 2018 at 9:45
  • 1
    $\begingroup$ @DavidHammen You are right, this was an incorrect oversimplification. I have reformulated the offending sentence. Let me know if you are happy with the correction. $\endgroup$
    – gerrit
    Commented Jul 29, 2018 at 20:49

Gerrit's got the technical answer; I'm going to answer for a layperson.

There are two ways objects lose heat.

  • The first, and the way people are most familiar with, is conduction. Something touches something else, and the hotter material transfers some of its heat to the colder material. It's why you rapidly lose heat if you wade into cold waters: your body is hot and the water is cold - so the energy from your body seeps into the water you're touching.
  • The second, though, is a bit harder to intuitively grasp: radiation. Or if it helps, think of it as light. Everything emits radiation/light - and that light is based on how hot the material is. Ever wonder why every material you heat up really how starts to glow red first (ovens, stoves, campfire embers, etc)? That's because when its that hot, the energy it's giving off is energetic enough to be seen by the human eye as red (the lowest energy color in the spectrum.) Heat it up further, and the color eventually shifts to a white.

Make sense so far?

Okay, Earth doesn't lose heat to conduction. Simply put, it's in a near-vacuum; there's not really anything its touching to transfer its heat to.

Instead, it loses its energy to radiation.

Now, objects emit radiation based on their temperature. So the energy coming from the sun? It's based on the temperature of the surface of the sun: about 6,000 Kelvin (about 10k degrees F). So the light/energy/radiation coming from it is very high-energy. That energy strikes earth and heats it up. But the Earth itself is comparatively cool: an average of 287 Kelvin (56 degrees F.) So the radiation that earth emits is much lower in energy.

And that's the key. Gasses like Carbon Dioxide don't allow lower-energy radiation to pass as easily has higher-energy radiation. So while it lets a lot of the sun's energy in, it doesn't let the radiated energy back out as easily.

  • 3
    $\begingroup$ What you call transfer should better be called conduction. Radiative transfer is also a thing, and it's what you call simply radiation. $\endgroup$
    – Ruslan
    Commented Jul 29, 2018 at 21:09
  • 3
    $\begingroup$ In addition to what @Ruslan says, conduction, convection, and radiation are all methods for heat transfer (in a way, so is advection). So the word transfer is indeed poorly chosen here. $\endgroup$
    – gerrit
    Commented Jul 29, 2018 at 22:47
  • 1
    $\begingroup$ Even out on a cold winters night the majority of the heat transfer from you will be radiative, unless it's very windy and/or you're wet. $\endgroup$
    – caf
    Commented Jul 30, 2018 at 1:22
  • 1
    $\begingroup$ Edited based on comments. And I learned something new - I didn't know that, even on a cold night, the majority of heat loss was still radiation. So I changed it to another common experience - being in cold water. $\endgroup$
    – Kevin
    Commented Jul 30, 2018 at 13:27
  • 1
    $\begingroup$ For a layman's explanation, I was surprised you used Kelvin instead of Celsius. And what's with the "10k degrees F" instead of "10,000 °F"? $\endgroup$ Commented Jul 31, 2018 at 14:15

To add to Gerrit's excellent answer, I'd like to add a couple more Images. Images always help clarify things for me.

Firstly, this one shows the spectrum light coming from the sun in red. The peak is in the visible range*. It also shows the thermal radiation from the earth in blue. This is in the infrared range. Below, it shows how different gases allow transmission of light at different frequencies. There's a lot going on, so for now, just pay attention to the row for Carbon Dioxide - the most well known greenhouse gas. Note how it allows close to 100% transmission in the visible range, but there is a gray patch in the range of frequencies where the earth is radiating. (I've added a circle around this patch.) Atmospheric Transmission [https://commons.wikimedia.org/wiki/File:Atmospheric_Transmission.png]

Secondly, there's this picture of a metal rod that's been heated to "white hot". At this temperature it is radiating in the visible range. The hottest part looks yellow-white (~2000K), the medium part looks red (~1000K), and the rest looks black (room temperature, ~300K). So you can see that objects with different temperatures radiate at different wavelengths. Red hot metal rod [https://commons.wikimedia.org/wiki/File:Blacksmith_working.jpg]

Actually, the black part of the rod is still radiating, but in the infra-red part of the spectrum, which our eyes can't see. You may be familiar with infra-red cameras being used to measure temperature of people, or as a kind of night-vision. So this last picture from an infra-red camera shows that even "cold" objects like people are emitting radiation. Thermal image - hand [https://en.wikipedia.org/wiki/SGR-A1#/media/File:Thermal_image_-hand-_1.jpg]

So to sum up, the sun is hot (6000K) and emits light, which is mostly in the visible range. Visible light passes through the gases in the atmosphere and heats the earth up to 300K. The earth emits infra-red light which does not pass through Carbon Dioxide. So Carbon Dioxide traps some of the heat which would otherwise have radiated away from Earth. More Carbon Dioxide traps more heat, causing the temperature of the Earth to slowly rise.

*saying "the peak is in the visible range" is the easiest way to understand it, but the causality is round the wrong way. Our eyes evolved to have the highest sensitivity in the range where the sun produced the most light.

  • 2
    $\begingroup$ Nice answer. The absorption spectrum of CO2 is actually the convincing thing. $\endgroup$ Commented Jul 30, 2018 at 9:10
  • 1
    $\begingroup$ That's the image I had wanted to add originally, but somehow I couldn't find it. Thanks for complementing my answer :) $\endgroup$
    – gerrit
    Commented Jul 30, 2018 at 9:44
  • 2
    $\begingroup$ Now I want an infrared image of someone standing outside in the winter cold, with one hand in a glove, one hand without a glove. The hand without a glove will appear warmer in the infrared image, yet colder to the owner of the hands. Although not an accurate illustration of the greenhouse effect overall, it may still be educational. $\endgroup$
    – gerrit
    Commented Jul 30, 2018 at 13:36
  • 2
    $\begingroup$ Just to clarify: on the first graph, the incoming/outgoing radiation is the red/blue line. The red/blue solid areas are what actually goes through once you take into account all the filtering and scattering (the grey areas in the graphes below). $\endgroup$
    – jcaron
    Commented Jul 30, 2018 at 17:13