🚀 go-pugleaf

In article <1992Aug2.062235.12886@ucsu.Colorado.EDU> knapp@spot.Colorado.EDU (David Knapp) writes:
>In article <1138@kepler1.rentec.com> andrew@rentec.com (Andrew Mullhaupt) writes:
>>Usually, one starts with Rayleigh scattering. This is in fact only an
>>explanation of why the sky is not red, yellow or green. It doesn't rule
>>out purple.
>
>It's Rayleigh scattering off oxygen, and that's about the end of it.

Are you sure? Let's take a peek in Jackson, _Classical Electrodynamics_
2d ed. p. 418:

	"Although Rayleigh's name should undoubtedly be associated with
	 the quantitative explanation of the blue sky, it is of some
	 historical interest that Leonardo da Vinci understood the basic
	 phenomenon around 1500. In particular, his experiments with the
	 scattering of sunlight by wood smoke observed against a dark
	 background ... anticipate by 350 years Tyndall's remarkably
	 similar observations."

Ok, Jackson is convinced. But in his explanation, we find that it is the
electron dipole moments which are important for "most gas molecules" - (which
seems to leave out your idea that oxygen is special) - But the punch line is
that the light is scattered more with the higher frequencies, and this
dependence doesn't stop at blue, either.

Now we have three things to do at this point. We want to claim that light
received in directions away from the sun is largely scattered sunlight,
(which is occasionally false, viz. the moon, Venus, etc.) but largely
true. I do not think that physics is mute on this point, either, but the
question you answer when you explain this piece of the blue sky story
_is exactly_ the explanation of why the sun is the brightest star in the
sky, (or in keeping with Olber's paradox, why it's dark at night). Jackson
offers _no explanation_ of this, but that doesn't answer any questions
which might arise. 

Then we need to fix the color of the sun. Rayleigh scattering on a planet
circling a hotter star than the sun would give a purple sky, and red
or even green sky for a cooler one. Jackson does not explain the color of
sunlight, but at least he plots the spectrum of the incident sunlight,
showing clearly that the sunlight doesn't have much violet to be scattered.

The final step is to deal with the possibility that more needs to be said.
Jackson deals with this by plotting the intensity of sunlight filtered by
a "dry, clean, isothermal, exponential atmosphere" and then remarking:

	"In reality the attenuation is greater, mainly because of the
	 presence of water vapor...ozone...as well as other molecular
	species and dust."

Water vapor and ozone have significant adsorption bands (outside the visible)
and it is fair to ask if the absorption in the visible is fully accounted for
by the Rayleigh scattering. In fact there _is_ some non-Rayleigh coloration
in the "sea level spectrum with sun at zenith" - it's not as blue as the
Rayleigh filter would predict (in Jackson's picture). There is no discussion
of what appears to be more absorption in the yellow or green not attributable
to Rayleigh scattering than any amount of absorption by Rayleigh scattering
in the visible.

Jackson may be quite correct that Rayleigh scattering explains the selective
absorption of sunlight by the atmosphere which we associate with the color
of the sky, but can one be blamed for thinking more can be said?

>>There is another objection to this - and that is, the explanation of why
>>the sky is blue must not be true at night. Nothing about scattering changes
>>at night, and so there must be a difference in illumination. Now there are
>>stars at night, and the sun is a star, so it is necessary that the proximity
>>of the sun is important. But if you consider a universe with three space
>>dimensions uniformly filled with stars, you can easily see that since the
>>number of sources scales exactly to offset the fraction of the light reaching
>>you, that the sky should be uniformly bright in every direction - and as
>>bright as the surface of the sun.
>
>Ignoring absorption, you have a serious point. But since absorption is quite
>real, this is a moot issue.  When you see a blue sky, you are seeing nothing
>but scattered sunlight.

>Well, since the intensity of all objects in the universe are dimished by 
>absorption by at least *something*, there is less luminous intensity at
>night. Like it or not. Therefore, there is no blue sky as detectable by the 
>human eye. Should be able to measure that rayleigh scattering from 
>stars with sensitive enough instrumentation? Yes.

Yes, and Rayleigh and Maxwell proposed measuring this scattered light as a
way to measure the number of molecules per unit volume, which apparently
succeeded (Jackson, p. 425).

So why isn't the sky blue when there is a full moon?

>Did I accidently get into rec.humor?!  I hope you don't think you can just 
>'see' doppler shifts of galactic objects with your bare eyes.

Actually, the idea is that you _might_ see Doppler shifts by _not_ seeing
starlight which has been red-shifted out of the visible. There is nothing
silly about this. However, I am given to understand that the Olber's
paradox is resolved by more than this red shift. You need to take into
account the finite age of the universe as well.

>>So far, I have not mentioned the sensitivity of the eye, as some posts
>>have. This is not a big factor, as we can see from the fact that when you
>>let sunlight into a dark room, the light does not appear blue - i.e.
>>the unscattered sunlight does not seem blue even though the differential
>
>Because that is the *unscattered* component.

You've missed the point, which is that you cannot explain the color of the
sky by claiming that it's a preferential response of the human eye. 

>It is the anti-stokes lines of Rayleigh scattering! Good grief! What more do
>you need? If you want to call it something else, go ahead, but study
>Rayleigh scattering and it will make more sense to you why the sky is 
>blue.

Still stand by this? Try explaining the blue color of the sky which occurs
at the zenith when the sun is setting using Rayleigh scattering.

Later,
Andrew Mullhaupt