Why Does the Sky Appear Blue During the Day?

[ZapperZ]

So you have to actually cite someone who had to undergo a drastic change such as this to being able to view UV frequency? Why stop there? Why not have someone's eye be imbeded with night-vision sensor and then claim that human can also view IR spectrum?

What gives?

Zz.

 

[Farsight]

I thought that would be something suitably definite and credible, that was so obviously not a speculation that you would be willing to countenance that people have a range of capabilities, and the 400nm cutoff is not absolute.

 

[Farsight]

And getting back to the original thread topic, saying "we live at the bottom of a rainbow" is an erroneous analogy. Other than pointing out that the blue in the sky is related to the bending of light in a rainbow, it does nothing to illuminate the user's understanding of the phenomemon. In fact, it steers them in an erroneous direction (as in my example) if they attempt to uindestand the analogy.

OK, hands up, I could have phrased it better. How about:

The sky is blue because of something called scattering, wherein blue wavelengths are absorbed by molecules in the air then re-emitted in a different direction, whilst longer redder wavelengths tend to pass unobstructed. This isn't the same as reflection where something has a definitive colour, and it isn't the same as a rainbow where refracted light is split into a spectrum, but has a similar effect of separating the light into different colours. These vary with the angle away from the direction of the sun. When you can look directly at the sun (at dusk or dawn) you see more yellow and red, and when you look away from the sun you see blue.

 

[ZapperZ]

OK, hands up, I could have phrased it better. How about:

The sky is blue because of something called scattering, wherein blue wavelengths are absorbed by molecules in the air then re-emitted in a different direction, whilst longer redder wavelengths tend to pass unobstructed. This isn't the same as reflection where something has a definitive colour, and it isn't the same as a rainbow where refracted light is split into a spectrum, but has a similar effect of separating the light into different colours. These vary with the angle away from the direction of the sun. When you can look directly at the sun (at dusk or dawn) you see more yellow and red, and when you look away from the sun you see blue.

This is still not exact, because if you say that it is the "similar effect of separating the light into different colors", where are the other colors? Why only blue get the preferential treatment?

No, it is specifically Rayleigh scattering, and how our human eye has evolved, that are the significant factors. And this has been done rigorously, and not simply via wishy-washy, handwaving argument. See

1. B.R. Holstein, Am. J. Phys. v.67, p.422 (1999).
2. G.S. Smith, Am. J. Phys. v.73, p.590 (2005).

Introducing diffraction into this explanation simply adds to the confusion, especially when it has no role in the explanation.

Zz.

 

[Farsight]

If anybody could offer a brief Why the Sky is Blue explanation that I could relay to a child, I'd be grateful. I'd also be grateful if anybody could post any links as to why the sky isn't purple. I found this on google, but didn't follow the explanation:

http://www.livescience.com/forcesofnature/050719_blue_sky.html

Two spectra that have the same cone response are called metamers. Smith stressed that this only concerns the neural signal coming out of the eye - long before any processing by the brain. "In previous research, people excised cones from the eyes of dead people and measured the response to different spectra," he said. "The same trick that makes red and green turn into yellow is happening in the sky. But in this case, the sky's combination of violet and blue elicits the same cone response as pure blue plus white light, which is an equal mixture of all the colors..."

 

[Hurkyl]

Your examples are not valid, simply because what you think as a direct observation of a "UV" light is really a secondary effect the same way that I am "viewing" my laser using business cards.

I don't think it's even that good -- AFAIK there isn't any secondary effect from UV light here... it's just a plain, ordinary visible light phenomenon that happens to lie below the violet band.

 

[Hurkyl]

The simplest description I've seen for why the sky appears blue is this:


The sky appears blue because air is blue.


All this talk about scattering light is analyzing too far -- the usual intent of the question is not asking for the underlying physical mechanisms that causes things to be colored, but instead expresses curiosity about the "fact" that air is colorless, and yet the sky is blue.

The easiest answer is that air, in fact, is not colorless, but is blue. But because the color is so faint, we only notice when we're looking through miles of air.

 

[Cyrus]

The simplest description I've seen for why the sky appears blue is this:


The sky appears blue because air is blue.


All this talk about scattering light is analyzing too far -- the usual intent of the question is not asking for the underlying physical mechanisms that causes things to be colored, but instead expresses curiosity about the "fact" that air is colorless, and yet the sky is blue.

The easiest answer is that air, in fact, is not colorless, but is blue. But because the color is so faint, we only notice when we're looking through miles of air.

I have not read this entire thread, so sorry.

I am not so sure about that Hurkyl. If what you said were true, then as viewed from space, we would observe the Earth to be primarily blue, and this is not the case. We can clearly and distinctly make out the Brown and Green terrain in contrast to the deep blue oceans. We do not observe a blue tint over land masses.

Further proof that scattering is the primary mode of why the sky is blue is due to the appearance of a red or orange sky at night. As the relative position of the sun changes, different wavelengths of light get scattered. This is why the sky is no longer blue, but red or orange.

...provided that I recall properly.

Over long distances, the air appears to wash out colors with a white or grey tint, not blue.

In the case of a Red sky, the angle of the sun is no long directly over head. The large angle results in the light having to travel through a much larger portion of the atmosphere. This causes the blue and violet spectrum to become lost, leaving the sky red, yellow, and purple.

As Zz already said, it is due to Rayleigh scattering.

 

[Doc Al]

All this talk about scattering light is analyzing too far -- the usual intent of the question is not asking for the underlying physical mechanisms that causes things to be colored, but instead expresses curiosity about the "fact" that air is colorless, and yet the sky is blue.

The easiest answer is that air, in fact, is not colorless, but is blue. But because the color is so faint, we only notice when we're looking through miles of air.

I think you are mixing up air and water; "why the sky is blue" with "why the oceans are blue". The sky is blue primarily due to Rayleigh scattering (as mentioned repeatedly in this thread). On the other hand, I believe that water is blue, not colorless.

 

[Farsight]

You run into problems with blue sea and white clouds, hurkyl.

I believe it's the same effect for both air and water, Doc:

http://www.madsci.org/posts/archives/feb98/887691022.Bc.r.html

cyrus: I'm told you can emulate the situation at home with a bright light shining thought a tank of very diluted milk, wherein you see blue from the side, and orange looking through the tank up into the beam.

Anybody: If I overlap a blue light with a violet light do I see light blue?

 

[Doc Al]

I believe it's the same effect for both air and water

I don't believe so. While air preferentially scatters blue light, water selectively absorbs red light.

 

[Hurkyl]

I am not so sure about that Hurkyl. If what you said were true, then as viewed from space, we would observe the Earth to be primarily blue, and this is not the case. We can clearly and distinctly make out the Brown and Green terrain in contrast to the deep blue oceans. We do not observe a blue tint over land masses.

I'm going by this explanation, which I particularly liked:

http://www.eskimo.com/~billb/miscon/miscon4.html#blu

What you say would only happen if the air was opaque, or had a very strong color.

Further proof that scattering is the primary mode of why the sky is blue is due to the appearance of a red or orange sky at night. As the relative position of the sun changes, different wavelengths of light get scattered. This is why the sky is no longer blue, but red or orange.

I never said that scattering is not the primary mode -- I'm saying that's simply analyzing the problem too deeply for the purpose of the question. Just like any other blue thing, air does something to light that causes us to see a blue color. Sure, the mechanism by which that happens is different for air, but I agree with the author of that page in that that's no reason not to simply say that air is blue.

 

[Cyrus]

I would like to point this out to you from his website:

They're wrong: you don't need complicated physics to understand this. The sky is blue for a very simple reason:

Air is not a perfectly transparent material. Instead it is blue!

Not quite. It may be true that air actually has some 'tint' to it; however, this is not the primary mode of why the sky is blue.

The sky is blue partly because air scatters short-wavelength light in preference to longer wavelengths. Where the sunlight is nearly tangent to the Earth's surface, the light's path through the atmosphere is so long that much of the blue and even yellow light is scattered out, leaving the sun rays and the clouds it illuminates red, at sunrise and sunset.

Scattering and absorption are major causes of the attenuation of radiation by the atmosphere. Scattering varies as a function of the ratio of the particle diameter to the wavelength of the radiation. When this ratio is less than about one-tenth, Rayleigh scattering occurs in which the scattering coefficient varies inversely as the fourth power of the wavelength. At larger values of the ratio of particle diameter to wavelength, the scattering varies in a complex fashion described, for spherical particles, by the Mie theory; at a ratio of the order of 10, the laws of geometric optics begin to apply.

Individual gas molecules are too small to scatter light effectively. However, in a gas, the molecules move more or less independently of each-other, unlike in liquids and solids where the density is determined the molecule's sizes. So the densities of gases, such as pure air, are subject to statistical fluctuations. Significant fluctuations are much more common on a small scale. It is mainly these density fluctuations on a small (tens of nanometers) scale that cause the sky to be blue.

I see what you are saying Hurkyl. The sky does have a component of blue that is inherent to it; but, I am saying that this is not the reason why the sky is blue and that his analogy of stacks of air making the sky blue is incorrect. If you extend his simple premise and continue it to its conclusion, how do you account for red sunsets? You simply can't.

Morever, he does not claim what you are saying, that scattering is the primary mode. He is saying the opposite, that scattering is NOT a mode.

That's flat out wrong.

ref-1:http://www.eskimo.com/~billb/miscon/miscon4.html#blu
ref-2: http://en.wikipedia.org/wiki/Diffuse_sky_radiation

 

[DaveC426913]

Why the sky is blue - for the layperson. Submitted for your criticism:

The light from the Sun, though appearing white, actually contains many colours, including blue. Normally light travels straight. If the Sun's rays are not aimed directly at us, we would not see them at all. This is why space outside a spaceship is black, even when the Sun is shining.

So, because light rays normally travel straight, standing outside our house, the only light we *should* see would be coming *directly* from the sun, or bounced off nearby objects. The sky above us *should* be black because rays from the Sun *should* pass right over us, and so we wouldn't be able to see it at all. This is why the Moon's "sky" is black, even when the Sun is shining.

BUT, here on Earth, our atmosphere scatters some of the light, sending it in other directions than straight. It scatters reds the least, and blues the most. So, some of the light that *should* have gone right over our heads, missing us completely, actually is bent, and ends up hitting us. This happens more with blues than with other colours of the Sun's light.

When we look up at the sky, we are seeing rays of light (mostly the blue ones) that have been bent towards us by the scattering effect of air.

 

[russ_watters]

I am not so sure about that Hurkyl. If what you said were true, then as viewed from space, we would observe the Earth to be primarily blue, and this is not the case. We can clearly and distinctly make out the Brown and Green terrain in contrast to the deep blue oceans. We do not observe a blue tint over land masses.

No, for the same reason why sunspots appear black: They are slightly cooler and dimmer than the rest of the sun, yet if the entire sun were the brightness of sunspots, it'd still blind you. They just get drowned out.

So the sky is bright enough against the backdrop of space to appear blue and not bright enough against the brightly lit Earth to make it appear blue from space when looking directly at it.

Note, though, that on an extremely clear day, the sky looks much darker the closer to the zenith you look because you are looking through less atmosphere.

Note also that looking through much more atmosphere from space (like on the edges), the Earth does take on a blueish tint, ie in South America (not to mention the atmosphere itself above Brazil) in this photo: http://visibleearth.nasa.gov/view_rec.php?id=2429

 

[Cyrus]

Quite right Russ. I did not consider that effect.

Around the edges of the picture, you can see a layer of the atmosphere in it's blue glow. Are you supporting that this blue glow is due to the scattering of light from the atmosphere, or that it is the color of the air due to 'thickening'?

 

[Gokul43201]

The simplest description I've seen for why the sky appears blue is this:


The sky appears blue because air is blue.

...except at dusk and dawn, when the air is either yellow, pink, or orange.

(So in the specific context of the question - being during times when the sky is blue - the answer provided by Hurkyl is something of a tautology.)

 

[Farsight]

I think that's pretty good Dave.

 

[russ_watters]

Around the edges of the picture, you can see a layer of the atmosphere in it's blue glow. Are you supporting that this blue glow is due to the scattering of light from the atmosphere, or that it is the color of the air due to 'thickening'?

Scattering. And...

...except at dusk and dawn, when the air is either yellow, pink, or orange.

If you mean the atmosphere itself (as opposed to the clouds in the atmosphere), no. No amount of thickness will make it appear red/orange, otherwise the horizon would always appear red/orange. What happens at dawn/dusk is that the sun/moon shine through/reflect off of thin clouds/haze and the atmosphere scatters out the blue light due to the added thickness, leaving the couds/haze red.

As seen in The Swerve's award winning photo, it is the clouds, not the sky behind them that are colored: http://i22.photobucket.com/albums/b326/PaidVacation/NaplesSunsetredgold.jpg

 

[Hurkyl]

Are you supporting that this blue glow is due to the scattering of light from the atmosphere, or that it is the color of the air due to 'thickening'?

Both. The latter is the simple, straightforward explanation, and the former is the low-level underlying physical mechanism that causes the latter to happen!

 

[Gokul43201]

...otherwise the horizon would always appear red/orange.

Not necessarily true (I think the mean free path of a photon in the lower atmosphere is smaller than a meter, but I may be wrong). But I haven't thought about this enough to say anything intelligent...so I recant my previous objection for now.

Edit : The more I think about it (okay that's only 20 minutes so far), the more I think it is wrong to speak of the "color of a gas". The phenomenon that "gives color" to a gas is very different from the thing that gives color to a solid. With a solid (to be precise, an opaque solid), once you go beyond the nanometer size range, color is a material property. The color of a thing does not depend on how much of the thing there is. The phonon dispersion relation is size independent. So you can speak of the color of a solid without having to specifiy how much of the solid there is.

I don't believe the same is true of fluids - where the physics itself is completely different from that of solids. The color of a fluid is determined roughly by a Beer-Lambert relation with the extinction coefficient determined by something like the Rayleigh formula. That makes it seem plausible to me that the "color of a fluid" is not a material property.


Edit2 : Looking for a reliable opinion, I decided to see if Baez has anything to say about this. From his website:http://math.ucr.edu/home/baez/physics/General/BlueSky/blue_sky.html

Sunsets

When the air is clear the sunset will appear yellow, because the light from the sun has passed a long distance through air and some of the blue light has been scattered away. If the air is polluted with small particles, natural or otherwise, the sunset will be more red. Sunsets over the sea may also be orange, due to salt particles in the air, which are effective Tyndall scatterers.

Too handwavy, by Baez's standards but he seems to be saying that it is the sky that is yellow (though the wording is admittedly a little ambiguous to me). Also, his picture shows that it is the sky, not the clouds in front, that is colored.

Also, I think back about the colorimetry experiments I did in college, where you determined the concentration of a solution by matching colors with a standard solution in a column where you could adjust its height. The color of the standard solution is a function of the height of the liquid column, though admittedly, a blue solution never turned red by making it a few inches taller.

Looking back through the thread, I see Zz has linked a couple of AJP articles. Better read those first.

 

[Farsight]

Oh hurkyl. Next you'll be telling me oil is rainbow-coloured!

 

[Hurkyl]

The more I think about it (okay that's only 20 minutes so far), the more I think it is wrong to speak of the "color of a gas". The phenomenon that "gives color" to a gas is very different from the thing that gives color to a solid.

I agree that they are different phenomena, but given the diverse use of the word "color" in everyday speech, I don't think it would be reasonable to restrict the word to mean nothing but diffuse reflection.

Oh hurkyl. Next you'll be telling me oil is rainbow-coloured!

I wasn't planning on it. The scattering we see in the air and diffuse reflection off of a solid have roughly the same bulk effect on light -- they select certain frequencies of light, and diffuse them. These two effects are far more similar than anything else we might call "color", so I think I'm justified in the use of the term.

Thin film interference is not a similar effect.

 

[russ_watters]

Edit : The more I think about it (okay that's only 20 minutes so far), the more I think it is wrong to speak of the "color of a gas". The phenomenon that "gives color" to a gas is very different from the thing that gives color to a solid.

Well, I was going to comment on Hurkyl's tautology before, but I guess now would be a good time: So what? So what if the mechanism by which different things cause a certain wavelength of light to get to your eye is different? Why does that make it wrong to say (for example) that hydrogen is red because of its emission spectrum while a brick is red because it absorbs everything else while reflecting red light. Sure, Hurkyl's tautology doesn't explain the mechanism (and there are several), but it most certainly is correct.

With a solid (to be precise, an opaque solid), once you go beyond the nanometer size range, color is a material property.

Who says that only solids that reflect and absorb are allowed to have a property called "color"?

The color of a thing does not depend on how much of the thing there is. The phonon dispersion relation is size independent. So you can speak of the color of a solid without having to specifiy how much of the solid there is.

That's correct, but so what? I guess you are saying that since it takes more atmosphere to be bluer, it isn't an innate property. But the same applies to solids: slice any solid thin enough and it starts to lose its color as well. So you can still say the same about a gas - not to mention a translucent material like the plastic wrap I make red filters. I fold it several times to make it block enough light to be useful on my laptop. So what if it is only slightly red at 1-ply? It is still red.

I have some filters for my telescope that are extremely high transmittance color filters. Because of their transmittance, you cannot tell what color they are just by looking at them (they completely transmit one color and completely absorb or reflect the rest). So the only way to tell the color is to hold them up against a white piece of paper or white light. But I guess for the sake of quibbling, we could say the red filter is red because it transmits red or blue/green because it reflects blue/green. But either way, the color is an innate property of the material.

I don't believe the same is true of fluids - where the physics itself is completely different from that of solids. The color of a fluid is determined roughly by a Beer-Lambert relation with the extinction coefficient determined by something like the Rayleigh formula. That makes it seem plausible to me that the "color of a fluid" is not a material property.

What determines the Beer-Lambert relation? Is it the same for all fluids. It seems pretty self-evident that if it depends on the properties of the fluids and is unique for each, then it can be said that the resulting color itself is a property.

Edit2 : Looking for a reliable opinion, I decided to see if Baez has anything to say about this. From his website:http://math.ucr.edu/home/baez/physics/General/BlueSky/blue_sky.html

Too handwavy, by Baez's standards but he seems to be saying that it is the sky that is yellow (though the wording is admittedly a little ambiguous to me). Also, his picture shows that it is the sky, not the clouds in front, that is colored.

You're talking about this:

When the air is clear the sunset will appear yellow, because the light from the sun has passed a long distance through air and some of the blue light has been scattered away. If the air is polluted with small particles, natural or otherwise, the sunset will be more red. Sunsets over the sea may also be orange, due to salt particles in the air, which are effective Tyndall scatterers. The sky around the sun is seen reddened, as well as the light coming directly from the sun. This is because all light is scattered relatively well through small angles--but blue light is then more likely to be scattered twice or more over the greater distances, leaving the yellow, red and orange colours.

He's talking about the sun itself except where he specifically mentions the sky around it. "Sunset" = setting sun in that context.

 

[Gokul43201]

What determines the Beer-Lambert relation? Is it the same for all fluids.

The relation is just that - what differs is the extinction coefficients.

It seems pretty self-evident that if it depends on the properties of the fluids and is unique for each, then it can be said that the resulting color itself is a property.

The relation tells you that different colors have different intensities that depend on the type and amount of fluid that the light passes through. This makes the color of air dependent on both the nature of the molecules and the amount of air along the path of light.

 

[DaveC426913]

No, it is specifically Rayleigh scattering, and how our human eye has evolved, that are the significant factors. Zz.

What does evolution of the human the human eye have to with anything?

Is there some suggestion that
- animals (those with colour vision) do not see a blue sky?
- cameras do not see a blue sky?

If there were no humans (or animals, or cameras) on the planet, the blue wavelengths of light would still preferentially reach the ground.

 

[Farsight]

Funny thing colour. It's only in our heads. What's real is the frequency. Maybe. Anyhow, I think it's to do with our Red/Green/Blue receptors:

http://www.exploratorium.edu/snacks/colored_shadows.html

http://en.wikipedia.org/wiki/Eyeball

If I shine a blue spotlamp on to a screen and overlap it with a purple spotlamp, I wonder what colour I see? I must try this out.

 

[Claude Bile]

What does evolution of the human the human eye have to with anything?

Is there some suggestion that
- animals (those with colour vision) do not see a blue sky?
- cameras do not see a blue sky?

If there were no humans (or animals, or cameras) on the planet, the blue wavelengths of light would still preferentially reach the ground.

Not true, violet and UV would preferentially reach the ground. In order to explain why we observe the sky to be blue, you need to take the colour responses of the human eye into account.

Cameras see a blue sky, because their RBG algorithms are designed to mimic the response of the human eye. A much more objective measurement would be with a spectrometer, such a measurement clearly shows significant amounts of violet and UV radiation scattered toward the surface in addition to blue.

I think this paper may be of interest to some (also contains the spectrum of the daytime sky I referred to) -

"Human colour vision and the unsaturated blue colour of the daytime sky" Glenn S. Smith, American Journal of Physics, 73 (7), 590-597 (2005).

Claude.

 

[vinter]

hey hey! can someone answer this please - why is the sun red in the evening? isn't THAT also because of scattering? why does scattering sometimes prefers red and sometimes blue?

 

[Farsight]

The scattering means the shorter bluer wavelengths get scattered off sideway more than the longer redder wavelengths.

So when you look up at the midday sky, you're seeing blue wavelengths scattered to one side. When you look straight at the sun in the evening through a lot of air you see more of the longer redder wavelengths because the blue stuff get scattered off to one side before it got to you.