Why aren’t sunsets red when viewed from low earth orbit? So “Why are sunsets red?” seems to be a pretty common question and it has to do with Rayleigh scattering in the atmosphere. I’m following some game dev tutorials (example implementation) on how to model and simulate this effect in computer graphics with the relevant scattering equations.
Every model has a strong red tinge near the terminator, especially when viewed from the night side of the earth, looking towards the sun, much like how sunsets look from the ground.


Yet, when I look at various NASA images from the ISS, the red color is barely visible at all!



This is the only image I could find that contains even a trace of red coloring.
All atmospheric scattering implementations I have seen use the line integrals from the nvidia article linked earlier. What are they missing?
 A: In ISS, the red phase of sunset, that is the descent of the Sun through the troposphere, lasts only a few seconds. In contrast, the blue crescent of the upper atmosphere is visible for several minutes. Therefore the red phase is missing on many sunset photos from ISS. 

Sunrise at ISS. Source: NASA/Scott Kelly
Similar photo and video
A: There's some interesting info in a more recent paper on this topic: E. Bruneton and F. Neyret, Precomputed Atmospheric Scattering. This paper, among other things, cites O'Neil's paper and discusses some of its (and other papers') shortcomings.
Also, although this new paper is an improvement as compared to O'Neil's paper, it still has its limitations, which the authors did note:

Limitations A limitation of our method is that the aerosol
  properties are assumed constant, depending only on altitude,
  whereas in fact they can greatly change depending on the
  atmospheric conditions [Slo02]. Since our precomputations
  are very fast we can change these properties quickly, but they
  remain uniform.

This might be one of the reasons, i.e. that they take aerosols to have some particular altitude distribution everywhere, while it may vary much. If aerosol density is exaggerated, you'll get the too high reddish component.
One of the good things about this paper is that its demo is available on GitHub, so you can download it and play with parameters (including source codes). Online demo is now also available.
A: The red sunset is due to Mie scattering. Gustav Mie, a not so well known physicist who had early ideas similar to gauge theory, worked on light scattering by micron-sized particles. The physics is rather complicated for classical EM. The result is that the intensity of scattered radiation of wavelength $\lambda$ by spherical particles of radius $r$ a distance $d$ from the observer is
$$
I~=~\frac{I_0}{2d^2}\frac{d^6}{\lambda^4 r^6}(1~+~cos^2\theta).
$$
This has a curious consequence, in that for $\theta$ small the scatter is large, but also the case happens for $\theta$ near $\pi$ or $180^o$. The next time you see a great sunset look east and you might find there is a red sky there too. This is the Mie backscatter.
