Consider the Great American Eclipse (The total solar eclipse over America in 2017). Observed atmospheric temperature drops were in the range of 3 C to 8 C in a matter of minutes. Surface temperatures dropped about the same amount as did the atmosphere, best I can tell. I'm lacking data on that though.

Other total solar eclipse observations universally (to my knowledge) document temperature drops.

In any event, such temperature drops represent a massive amount of energy presumably lost to space in a very short amount of time.


How does this energy in the atmosphere radiate to space?

  • The Energy must be radiated to space
  • Greenhouse gases can absorb kinetic energy from O2 and N2
  • Only greenhouse gases can radiate at atmospheric temperatures
  • Without solar input, the atmosphere cools because greenhouse gases radiate more energy than the earth can supply

Is that correct?

  • $\begingroup$ There must be a massive energy flux out to space, to balance the massive energy influx from the Sun. When the latter is suddenly removed, the balance shifts suddenly $\endgroup$
    – RC_23
    May 22 at 16:15
  • 1
    $\begingroup$ I’m not quite sure what the question is here. Are you asking whether the mechanism for temperature drops during an eclipse is different from the mechanism for making it get colder at night? $\endgroup$
    – rob
    May 22 at 16:58
  • $\begingroup$ Energy transfer can only occur via conduction, convection, or radiation in the atmosphere. The only gasses capable of significant transfer of energy in the atmosphere are greenhouse gases. O2 and N2 radiation at atmospheric temperatures are negligible. In cool dry air, the only significant greenhouse gas that could be cooling this body of air that fast is CO2. Can CO2 actually cool the atmosphere? How does that square with CO2 trapping heat? $\endgroup$
    – PaulSnow
    May 23 at 15:22

1 Answer 1


Water is the main GHG in the terrestrial atmosphere. But that's besides the point. The surface air temperature is set by the balance of visible transparent irradiation and diffusive infrared re-radiation.

However, Earth's infrared optical depth due to all the Greenhouse gases is only $\tau_{\rm IR}\approx2$, so it is not surprising that a cooling effect can happen in minutes via radiation. The infrared radiation simply diffuses away once it is not resupplied anymore.
Same as if it were suddenly to become night.

  • $\begingroup$ Also, I don't understand the τIR≈2. Various sources give far lower values. Diamond is supposed to be the highest at 2.42... $\endgroup$
    – PaulSnow
    May 24 at 5:07
  • $\begingroup$ @PaulSnow Without greenhouse gases, the atmosphere would be transparent to the outgoing infrared radiation, and hence cool maximally. It would reach radiative equilibrium with the surface only at $T=T_{\rm eq}$. When adding GHG, your cooling surface, which cools with the power $L = 4\pi r^2_{\tau=1} T^4_{\tau=1}$ from the $\tau=1$ transition, will move up the adiabatic gradient towards lower $T_{\tau=1}$ (while $r_{\tau=1}$ only changes very little), hence decreasing L, which means that radiation must be left behind. This is the greenhouse effect. $\endgroup$ May 24 at 9:13
  • $\begingroup$ @PaulSnow Not sure what you're talking about. To avoid confusion I specifically said "the infrared optical depth". Look up the definition if you're unsure. But essentially, when computing the integral over densities $\rho$, path length r and opacity of the gas mix $\kappa$ you end up with $\tau= \int \rho(r) \kappa(r,T,\rho) dr$ of approx. $\tau \approx 1.93$ for a saturated steam water atmosphere on Earth, and $\tau\approx 1.98$ when adding ~300ppm CO2 into the mix. This $\tau$ is enough to raise the temperature above $T_{eq}$ by a few tens of K, but also the diffusion time is not long. $\endgroup$ May 24 at 9:18
  • $\begingroup$ I understand roughly the radiation component of warming the earth, but I struggle with the conduction and convection components. On an earth with atmosphere of only O2 and N2 and other non-greenhouse gases, the atmosphere would still absorb heat from the surface during the day, but would be unable to shed that heat at night through IR radiation, and be too cool to shed that heat via higher bandwidths. Since conduction only involves a few cms of air at the surface, heat in the atmosphere would be trapped. $\endgroup$
    – PaulSnow
    May 30 at 13:48
  • $\begingroup$ It also seems to me that without CO2, calm dry air under an eclipse could not cool. Am I wrong about that? $\endgroup$
    – PaulSnow
    May 30 at 13:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.