The colors of green flashes depend primarily on the clarity of the air. It's now possible to simulate these colors fairly realistically, including the effects of aerosol attenuation and scattering, as well as extinction by the molecular atmosphere.
Here are some examples, all showing the same inferior-mirage flash, but with different amounts of aerosol haze. The simulated images show just the zone of sky between the astronomical horizon (at the top of each image) and the apparent horizon, as seen from 2 meters above the sea. The same atmospheric thermal model is used for all; only the vertical aerosol optical depth (indicated by the numbers across the top of the table) and the cutoff wavelength are varied.
The cutoff wavelength is the wavelength of light that has just disappeared at the center of the flash; that is, only shorter wavelengths contribute to the flash you see. These wavelengths are listed at the left side of the array of images. The cutoff is constant along each row; time (at sunset) progresses down each column.
The haze scale height is taken to be 1 km. Average clear conditions at sea level might correspond to an aerosol optical depth of 0.05 (the rightmost column of the array of images). Exceptionally clear conditions at sea level might by 0.01 or 0.02; anything less would be extremely rare. Lower aerosol opacity is common at considerable elevations (in the mountains), but not at sea level.
Each image is scaled to make it as bright as possible without saturating any pixel. That makes the sky appear to get brighter as you go down each column, which is the direction of time for a single sunset. Don't be fooled: each column of images represents only a couple of seconds of time, during which the sky brightness would not change perceptibly. What really happens is that the sky stays nearly the same, while the flash rapidly fades away. The faintest images are brightened up here, so you can see the dim flash as it disappears.
Wavelength Aerosol opacity →
Considering that an aerosol opacity of 0.05 is typical of “clear” conditions at sea level, you can see why green flashes aren't ordinarily observed: that (extreme right) column of the array shows only yellow-orange flashes, because too much red sky light is superimposed on the flash. But, for very clear conditions (0.01 or 0.02), a nice green (but not blue) flash is visible.
In these conditions, there isn't much change in color of the flash as it disappears. The reduction in cutoff wavelength with time should make the green flash become bluer; but the increasing relative contribution from the sky holds the perceived hue nearly constant for a second or so.
As you can see, blue and even “violet” flashes are possible, but only under extremely clear conditions. That's also in agreement with observation. The limiting aerosol opacity for a blue flash seems to be between 0.002 and 0.005 — and even these pass through the blue stage so briefly that their detection might be marginal.
However, in the mountains (the left three columns of the array), blue and even violet flashes are visible. And here, the change in color in the earlier stages is plainly visible, because the fainter (and bluer) parts aren't drowned out by the sky.
The top row of the array shows a green flash with a visibly yellow core. This is the view that would be obtained with binoculars; the resolution of these images is 1/8 of a minute of arc, while that of the eye is about one minute.
However, if you stand far enough away to see the flash as just an unresolved spot of light, its average color is more or less green (in that top row). So let's consider that as a reference point, and count seconds after the disappearance of 600-nm light. Here are the numbers:
|wavelength||depression||(at Equator)||(45° lat.)|
As you can see, ordinary green flashes at low latitudes should be visible about 2/3 of a second. But observers usually report between 1 and 2 seconds. The additional duration usually perceived at sunset must be due to the perception of a green hue before nominally green light is dominant; this effect is due to retinal bleaching.
You might like to compare these computed colors with the observed colors of green flashes.
Copyright © 2005 – 2006 Andrew T. Young
GF pictures page
main mirage page
GF home page
or the website overview page