After seeing a particularly vivid rainbow the other day, it occurred to me that it would be an awfully neat coincidence if the full range of a rainbow corresponded to human colour vision. Is there anyway of know how far out into the UV or IR (or even beyond, in either direction) a rainbow would display, if we had the ability to see it?
Since water and air are reasonably transparent to near UV and IR, the rainbow will extend past the visible spectrum on both sides. In fact, Herschel used a prism to discover Infrared light - presumably he could have used a very obedient rainbow.
A rainbow is reflected sunlight, so it depends on how much UV and IR the Sun emits. It turns out that the Sun emits most of its radiation in the visible range, which is probably why our eyes evolved to detect those wavelengths. The Sun does emit a lot of UV and IR as well, but most of it is absorbed by the atmosphere and doesn’t make it to the surface. So a rainbow does extend beyond the visible for a bit, but those colors would be much fainter even if we could see them.
True for near UV, not true for near IR. Sunlight - Wikipedia
Near IR would be nearly as bright as the other colors.
Reflected and refracted. Without refraction, you don’t get the splitting of colors.