Celestia Forums

What would the colour of the sky be for an Earthlike planet orbiting a red dwarf star?

I couldn't find this information anywhere, so I had to calculate the colour of the sky from first principles, and make a lot of guesses.

There are three things to consider:

• The light output of the red dwarf star.
• The scattering of light in the atmosphere.
• The frequency response of the human eye.

The light output of the red dwarf star is a blackbody curve, and red dwarf stars don't output much blue light.

Atmospheric scattering preferntially scatters blue light, with the amount of scattering proportional to the fourth power of frequency (thus 400 nm blue light is scattered 16 times as much as 800 near infrared light).

The frequency response of the human eye has to be taken into account so that the actual colour of the sky as perceived by humans can be determined.

My results suggested that the colour of the sky for such a planet would not be blue at all. With the red dwarf outputting so little blue light, there is not enough blue light available to make the sky blue. What I found was that the sky would be a pale colour that is more whitish than bluish. The pale colour could be a greenish white or even a yellowish white.

These results are not authoritive, because I have no way to verify them. Can anyone who knows something about optics tell me how accurate these results are?

I think you need to plot the product of the visual effects -- wavelength vs signal strength. That'd show the input signal, anyhow.

Unfortunately "how it looks" is going to include more than just the frequency response of the eye. Don't forget that there's a psycho-visual "white balance" effect which makes illumination seem to be colored more uniformly than it really is. Rooms lit by interior incandescent lighting seem to be quite white until you compare them with sunlit areas, for example. Then you discover that they're actually quite red.

We've had this debate several times, Grant suggests that the sky might be grayish, as the Rayleigh scattering tends to cancel out the slope in the wavelength distribution, see this post.