Why is Earth blue? (Synopsis)

“Whenever I gaze up at the moon, I feel like I’m on a time machine. I am back to that precious pinpoint of time, standing on the foreboding — yet beautiful — Sea of Tranquility. I could see our shining blue planet Earth poised in the darkness of space.” -Buzz Aldrin

If you look at Earth from space, you'll find that we're a blue planet. You might chalk that up to the fact that our sky is blue, the sky is the outermost layer of our planet, and hence the planet appears blue. But then why do the continents and clouds appear to be such different colors, and why is the "blue" of the ocean such a different shade from the sky we see?

Image credit: NASA’s Earth Observatory / MODIS. Image credit: NASA’s Earth Observatory / MODIS.

It's because the "blue" we see from space -- or the blue we see from looking at the ocean -- actually has nothing to do with the contents of our atmosphere!

If you ever thought the ocean was blue because it reflects the sky, you've got to read this.

More like this

Are those pictures the natural color, or have the colors been enhanced?

Note, if the ocean was blue because it reflects the sky, Greenland would also be blue because it reflects the combination of the sky plus direct sunlight. So apparently the
overall lighting isn't far from white.

By Omega Centauri (not verified) on 22 Sep 2015 #permalink

I seem to recall this coming up as a trick question on the TV show QI... asking "why does the ocean appear blue", the answer was "because it *is* blue".

@ OC#1
Winter photo - look at the north polar regions as well.

A trick question just occurred to me:

"What is the average color of the sky?" - with various images to select from. The trick is that most people might forget it's black at night, making for a very dark average.

It's not purple mostly because it's lacking "all" green and red. Since we have three colour receptors and the only one left is blue, we see it as blue.

If you don't believe it, go and check the midpoint of blue and look at a colour chart with that colour blue and see if it matches the sky.

It won't.

Because it isn't a single colour, it's the result of a spectrum. One that lacks red and green.

Oh, and the colour balance done automatically isn't in the eye. It's in the brain.

And it's even per-eye.

Perhaps you might want to look at a full spectrum camera shot of limited at just under the UV spectrum and try again WOW.

Last comment went into hyper space so lets try again, most good physics forums get it right


A full spectrum camera image with a cutoff frequency just under the UV comes up most certainly purple. Hit google and you will find many photographs from such cameras and the world is very purple.

If you look at any camera shot of a rainbow you will also see a vivid purple band but looking at it in the flesh most people only ever see down to blue yet if you search for an image on the net of a rainbow I assure you the camera will see the purple.

This one is for you WOW


The spectral line of the 3rd image is funny you wont see the purple yet if you grab a camera or video recorder and look at the image they will process the purple (they correct it for you) and it will show you it. The purple is actually in the image you just can't see it.

Purple is really weird sometimes you see it sometimes you dont :-)

"The purple is actually in the image you just can’t see it."

Look up the word Gamut.

We have to present it as a mix of blue and red in inks.

Correct but that is ONLY TRUE to humans. Google mantis shrimp or even some birds like peacock and look at there eye sight they see much wider color range and dogs/cats see much lower.

So the sky is only blue to the HUMAN's

"Correct but that is ONLY TRUE to humans."

Well excuse me for making that assumption! I only said "our eyes" and I'm human. Maybe you aren't.

So what are you, LdB? A bee? Squid? How did you manage to learn the English language?

You are sort of missing the point, what we are trying to get at is "spectral color" versus "mixed perceived color".

Spectral yellow is what 575nm or it can be represented to a human as red/green mix, we see those two things as yellow.

Pink for example does not exist as a spectral color the human mind makes it up when it sees red and blue mixes. There was a humor joke to have it removed as a color a while back or call it minus green :-)

Violet/purple exists as a spectral color but the human eye doesn't see spectral purple it only sees purple when presented in the mixture form.

I agree the sky is blue to our eye but from a spectral view it's actually violet/purple. Ethan's post starts why is the sky blue? That is a loaded question because he never deals with "from a human eye" which if you look was my very first comment was what it says. I simply commented Ethan missed dealing with the whole issue properly.

The sky is purple/violet from a spectrum point of view and any spectral analysis or an image from a full spectrum camera will show you that. The human eye doesn't see that color in it's spectral form so we see the next color which is blue.

So is the sky blue? Well it depends if I am looking at a spectrum analysis or with my actual eye and the answer is different. This is a science forum isn't it?

So I guess can I ask a more specific question WOW, if you know because I have never seen a spectrum of earth from space.

Is earth from space spectral blue or spectral purple and the later if I see from my human eye will be blue?

That in a nutshell wasn't clear to me from Ethan's post.

That is a similar argument to the pink is not a color one the one minute physicist turned it into a bit of a joke


There is no wavelength for Pink and a number of other colors it exists as only concept of qualia to an animal in it's mind.

That wasn't really what I was asking, the easiest way to probably ask the question is does anyone have an emission spectrum of Earth as seen from space. When I google it I get lots of climate science stuff and lots of arguments it because it's fairly complex as expected. I assume someone actually has turned a spacecraft around and done it from space.

On the physics forums this one says NASA was doing it but the link doesn't work


The nimbus mission data is very technical and so what I am really looking for is a simple emission spectrum graph of earth as seen from space, then there is no confusion about human eye rubbish.

"You are sort of missing the point"

You're making up an entirely new "point" so as to "win" and make me "lose".

The point is you don't KNOW what anything other than humans whose language you know and have met and asked the question what colour the sky is.

Asked your dog? When it went "Woof", did that tell you it things it's sort of woof-coloured? And did you know that colour was not blue but some sort of tan?

Heck, ask a Russian. He'll say the sky is tsvet siniy. Not blue.

Or is the point that he MEANS the word blue in our language?

In all cases, the point is we have to mean what human perception calls things when we call things what we see them as.

And gassing off about how blind molerats don't see the sky as blue, but would see it as brown if they could, really isn't the point.

"if you know because I have never seen a spectrum of earth from space."

I never have either, so why the hell ask me that specific question?

"Is earth from space spectral blue or spectral purple and the later if I see from my human eye will be blue?"

It isn't either, since spectral blue is a narrow range of colours, and spectral purple also a (different) narrow range of colours, whereas the earth is varigated. And even the blue is a wider band than "spectral blue" as defined by, for example, the XColor figure set mapped on to an incandescent display.

And any such spectral colour is received by the eye in a sensitivity pattern that changes based on the illumination and recent history of illumination on that eye, then processed through the vision centres with colour mapped to a set of qualia and it is THERE that our "colour" is made other than what it physically is.

We have limited words for colour, and blue is the one just about everyone would put for the clear daytime sky. If the sun were much much dimmer, and not a visible point source to make our qualia change, we may use the word purple. But the qualia purple in our language is almost always used for darker colours, or with significant red, whereas blue is more prevalent when there's bright light or a preponderance of blue and little else.

The night sky is blue, too, just so dark that we can't tell it from black. Starlight is just as good at being scattered by our atmosphere as our star's light is. There's a hell of a lot less of it.

SPECTRALLY you'll see a wide spectrum with some small fraction of 6000K thermal (white from recflected clouds and snow), 300K thermal (from the earth in the IR), some red, more orange and yellow (from the dirt reflection), a fair bit of green (from the plants) And blue (from the ocean), but each terrestrial object will display some other spectral reflection in the other bands too.

Put that on a screen as a single illuminance over the whole screen and you would call it "blue", in much the same way as we call the sun "Yellow". It's the closest single word to the colour we see from words in daily use.

If you went to the XColor codes you may pick out one of the cornflower colours as closer, but it won't be a name we usually use.

Oh, and remember, when it comes to vision,"red" will still excite some of the green receptors, where your eye has made the model "green" is. And green excites both the others. And blue excites green. The only way to get "prime" colour is to use a machine to measure it, calibrate the machine to a brightness of each photon frequency, then use a prism or diffraction grating to tease out the frequencies.

We don't use that in our visual system.

Long post to say you don't know WOW. There will be heavy emission in the far IR from earth I am sure but at the blue end I would guess from the sky behaviour that the emissions will be a bit of blue and lots of UV range. So my guess would be spectrally it will be loaded to violet which by eye we see as blue. That was all I was trying to work out and Ethan didn't address in the post at all.

I wasn't ever trying to win, the idea of winning is something only you care about. I had great fun taking you round in circles with the Lubos argument and trolling you because you actually need to win every argument.

You come across as a sort of a wanna be, try hard who doesn't have the IQ to really match it with the smart people like Lubos etc.

So basically after countless posts and a lot of hyperventilating we are back to what I first complained about with Ethan's article. None of us actually know what the color of earth looks like from space, we all just have guesses. For my part I wouldn't be shocked if I was wrong it is a complex thing.

Earth will look blue to the human eye at least in close which is no real surprise, but the wider more scientific answer we are no closer to.

"None of us actually know what the color of earth looks like from space"



I just said that I'd never BEEN in space to see myself. HOWEVER, I did explain what it would look like from space.

Are you just being a twat or are you genuinely ignorant?

"None of us actually know what the color of earth looks like from space"



I just said that I'd never BEEN in space to see myself. HOWEVER, I did explain what it would look like from space.

Are you just being a twat or are you genuinely ignorant? Because your post only makes sense if either of those are the case.

"Long post to say you don’t know WOW."

Nope, the bit saying i don't "know" by personal experience was one sentence.

TRY READING THE REST OF IT you ignorant moron.

No what you offered was a few guesses a lot of hand waving and not much else any actual data or facts to back any of your comments up? I already doubt you have an IQ worth taking seriously and then all you can do is trash talk and name call.

Sorry I don't believe you and there are good reasons because I actually read the papers on exoplanets and planetary emissions studies that have actually been done.

In fact there is quite some variation between on what has been studied with the violet and UV emissions on both earth and jupiter over there orbital movement. There is little pure visible spectrum data as there are limited space based platform capable of that.

In the end it is summed up by this statement from the paper "Spectra as Windows into Exoplanet Atmospheres"

"Despite multiple ground- and space-based campaigns to characterize their thermal, compositional, and circulation patterns (mostly for transiting giant planets), the data gleaned to date have (with very few exceptions) been of marginal utility. The reason is that most of the data are low-resolution photometry at a few broad bands that retain major systematic uncertainties and large error bars. Moreover, the theory of their atmospheres has yet to converge to a robust and credible interpretive tool. The upshot of imperfect theory in support of imprecise data has been ambiguity and, at times, dubious retrievals."

From what I have read and digested much of the thermal emission is masked by the atmosphere and as that isn't understood well, so the exact emission spectrum is a bit up in the air.

Now if you have actual data rather than trash talk and pure speculation and guesses please feel free to let me know.

So let me present the data I have been able to find. First there is the "pale blue dot" image from Voyager 1 and it's family image series.


You will however note the images are overlapping three different narrow bands of green, blue and purple and then there was a special mixing process.

"The majority of the frames were acquired in gray scale with the probe's Wide-Angle Camera, while the close-up views of each planet were acquired in color using the Narrow-Angle Camera."

This leaves open the problem that there may be much brighter emissions in other bands. As they state they selected bands at the time based on the atmospheric science state at the time.

"No what you offered was a few guesses"

No what I offered was exactly what you'd get as far as it can be written in language

Just because YOU'RE a fucking moron doesn't mean you have the right to dumb everything down to what you managed to make of it.

Ah yeah the tough guy trash talk again ... No data and no actual evidence then. We just have to trust the great WOW who knows everything ... LOL

It's okay WOW don't worry your huge IQ over it, I have asked on a real physics forum and might get actual real answers if they are known.

I thought given Ethan's background he might actually know the answer but he obviously doesn't read the comments.

For a time I actually wondered if the great WOW was nothing more than a sock puppet for Ethan to tackle dissent in the comments, because you got away with a lot against the religious fringe. I have gone off that idea you are just some poor lame thing that couldn't hold his own on a real physics forum and frequent this backwater.

Wow, are you going to let LdB talk smack to like that?

Get your ass over to that goofy Lubos blog and start kicking ass and taking names otherwise you loose credibility.
Heck he don't look like he weighs a buck 30 soaking wet, surely you can take him. Besides he's a Czech, they're below the Polacks on the brain food chain.

Now go get'er done.


By Ragtag Media (not verified) on 27 Sep 2015 #permalink

"Ah yeah the tough guy trash talk again "

Ah, the false identification to make u some inherent "problem" therefore nothing you need worry about.

Just because YOU'RE a retard doesn't mean everyone else is.

Even if Teabaggie here is doing his best to show how dumb people can be.

"It’s okay WOW don’t worry your huge IQ over it, I have asked on a real physics forum and might get actual real answers if they are known."

You did get an answer, but you don't want to know.

You'd rather clutch your pearls.

That is right I got the great WOW answer which he made up.

Yeah sorry I stick to science and data, I don't accept pseudoscience as and appeal to the great WOW authority as answers.

I have an idea that the answer is probably complex due to the atmosphere and so would like to see a measured spectra.

That's right, LdB, you're a moron. I didn't make it up, I EXPLAINED it.

You, meanwhile, haven't a frigging clue.

Gee wonder where I have seen your sort of explaination before google Pseudoscience.

"Pseudoscience is a claim, belief or practice which is incorrectly presented as scientific, but does not adhere to a valid scientific method, cannot be reliably tested, or otherwise lacks scientific status. Pseudoscience is often characterized by the use of vague, contradictory, exaggerated or unprovable claims, an over-reliance on confirmation rather than rigorous attempts at refutation, a lack of openness to evaluation by other experts, and a general absence of systematic processes to rationally develop theories."

Sounds a bit like how you explained it :-)

"In God we trust, all others bring data."

Look, LbD, we BOTH agree I'm smarter than you, and you are a dumbass.

You asked and were answered, but you prefer to whine instead.

Now you're bleating on about hard data, which, since the subject is the PERCEPTION of colour by the human visual systems, IS NOT POSSIBLE.

You have been told what the spectrum is: broad, with mostly blue dominating.

LdB, by a knockout.

But Wow is more fun to read.

True I will give you that he is funny to read and I wanted to know how smart he actually was so wasn't really helping :-)

What I really love now is the dog ate his data.

The reality is every one with a reasonable science background knows why the sky is blue and that opens up the problem.


The second image down shows the spectra at noon at the equator. Present that spectra to a spectrograph model and it will tell you it's violet but to the human eye will register as blue. That is just the way we see.

Lots of marine animals see well into the violet and even into the ultra violet for a very good reason. Some birds and other animals in jungle enviroment do as well for similar reason.

When I read Ethans question I actually wondered how much that spectra moved. There will definitely be less white and yellow which is to do with the suns behaviour in the atmosphere which again has a spectral reference shown here


The question I was actually pondering was if the spectra moved to the blue enough so the human eye and a spectrograph agreed on color which is the ultimate answer to the post question.

Strangely the spectra from space appears to never have been done, which I found sort of surprising. I guess if it has never become an issue then why do it.

If you look at the "pale blue dot" they shot green, blue, violet for obvious reasons if you look at that noon spectra. Strangely however the exact mixing they used for those sequences is not recorded, which was interesting. Carl Sagan was the one who pushed for the shots and used in his later book but there does appear to be some politics rather than science behind the mixing process to get "the right colors".

So ultimately there will be something that looks like that noon spectra which will be blue to the human eye. No real surprise there look at any photograph from space. The question as to whether a spectrograph will see blue I think is still out it could be anywhere from cyan to violet.

@Wow #36: No, LdB was not answered, certainly not by you. LdB made it clear that there was no misunderstanding about visual perception of color. The question LdB asked was specifically _not_ about visual perception, but about the underlying true spectrum, including very short wavelengths (UV).

LdB is quite correct that a naive analysis of Rayleigh scattering suggests something like a "UV catastrophe", where the physical spectrum should be dominated by the shortest available wavelengths, which are _not_ "blue", but rather UV. A more detailed analysis depends on the size of the features doing the scattering, the input spectrum, and so on.

The question, "What is the Earth's actual spectrum as seen from space?" is a very good one, and one which you, Wow, were unwilling to address, preferring instead to argue and insult. I do not know the answer to the question of the top of my head, either, but I am at least able to recognize a sensible, scientific question when I see one.

By Michael Kelsey (not verified) on 28 Sep 2015 #permalink

Thank you Michael I have a comment which obviously went into moderation and may pop up before your reply and I actually turned it to what a spectrograph would make of it.

In discussion on other forums and with experts that even causes problems even that isn't going to be exact but depend on the device and how it was calibrated .. max signal, bands etc. A couple pointed out that the whole question of what color the sky/earth although popular in science media (like Ethan asked) is actually really problematic. The sky as seen from the human eye is white + blue as a full spectra not some color, which I think we all agree and understand.

So I guess I complained at the start that Ethan had not covered or even mentioned the eye, where perhaps I should have complained about the question itself. So in some ways the answer I was seeking is as flawed as Ethan's answer.

I guess the really interesting thing I got out of the exercise was that we actually don't have an emission spectra of Earth to space which in itself was surprising.

@LdB #40: You wrote, "I guess the really interesting thing I got out of the exercise was that we actually don’t have an emission spectra of Earth to space which in itself was surprising."

I'm not entirely sure that should be surprising :-)

Earth-orbiting satellites are too close to get a sensible planet-averaged spectrum from observations. Off-planet probes are neither designed nor likely to waste fuel/power/bandwidth taking detailed observations of a planet for which we have much cheaper ways of getting detailed data.

The observations we do have of Earth from deep space (Pale Blue Dot, The Day The Earth Smiled, etc.) are really incidental, opportunistic results of other targeted probes. Even SDO (Solar Dynamics Observatory) is too close to the Earth to get a true "average spectrum".

An interesting exercise is to do things like use the Earth's reflection spectrum off the Moon as an simulation of exoplanet observations (search "Earth as an exoplanet" for papers), but I'm not familiar with broad spectrum results of that; the various studies have been fairly targeted toward water vapor, oxygen, etc.

By Michael Kelsey (not verified) on 28 Sep 2015 #permalink

I have been given a very precise way to rephrase the question which is very scientifically clear and even WOW wont be able to quibble with. What visible light band has the highest emission using the standard 6 band definitions

Violet 380–450 nm 668–789 THz
Blue 450–495 nm 606–668 THz
Green 495–570 nm 526–606 THz
Yellow 570–590 nm 508–526 THz
Orange 590–620 nm 484–508 THz
Red 620–750 nm 400–484 THz

From Dr. Christopher S. Baird this is the color spectrum of the sky considering Rayleigh scattering + thermal incident sunlight

So highest emission is violet (668–789 THz) .

So the question is which of the six bands is higher from space.

@Michael Kelsey your answer is exactly what I found and you are pretty much spot on. As you guessed I end up reading lots of exoplanet search papers trying to piece how all the spectra shift with atmospherics.

The moon idea is a really good one. On to that thank you for suggestion.

@Wow #36: No, LdB was not answered, certainly not by you.

Yes he was, Michael. Stop being a knob.

LdB made it clear that there was no misunderstanding about visual perception of color.

Yes he did. He went and jibbered off abut how OTHER CREATURES had different visual apparatus. Why? No fucker knows. Not even you.

The question LdB asked was specifically _not_ about visual perception, but about the underlying true spectrum, including very short wavelengths (UV).

Here you go, his first statement:

You actually missed part of the answer as to why the sky appears blue, Ethan. It works a you said the shorter wavelengths scatter more and the sky should appear purple to us not blue :-)

Please note: the sky should appear purple to us not blue

Given you have completely misunderstood what LdB is saying you should rethink your stance in the post #39.

LdB is quite correct that a naive analysis of Rayleigh scattering suggests something like a “UV catastrophe”

Where does he claim this? Nowhere. And who is DOING a naive analysis of Rayleigh scattering? And quantisation is no longer an abstruse topic. At least to anyone who knows about Rayleigh scattering and the UV catastrophe that made quantised light so necessary to explain (originally not by making light quanta, but by restricting the intensities of light at higher frequencies in a manner that LOOKS like quantisation).

the physical spectrum should be dominated by the shortest available wavelengths, which are _not_ “blue”, but rather UV.

Two problems AT LEAST:

1) We can't see UV, so it isn't visible to us, so why would be assign a colour to it?
2) The sun doesn't put out an increasing amount of radiation in the UV range, being a black body at 6000K rather than at 100,000K. There aren't the photons to make UV scattering dominant.

A more detailed analysis depends on the size of the features doing the scattering, the input spectrum, and so on.

Which is what I gave in post 18, roughly the middle bit, starting with SPECTRUM. Which LdB then whined about no existing, and you haven't bothered to read at all, preferring to leap to a conclusion that LdB must be right because I'm a meanie.

A PROPER scientist accrues DATA *before* making a conclusion. A quack leaps to one first.

Isn't that right, ducky?

The question, “What is the Earth’s actual spectrum as seen from space?” is a very good one, and one which you, Wow, were unwilling to address,

Given it wasn't LdB's original question, but answered later, rather secondary, really. Given too I have answered it already in post 18 your asinine accusation here is entirely due to your lack of perception with your head shoved right up your arse.

I do not know the answer to the question of the top of my head, either,

Neither am I, apart from the rather handwaving description, which is both more than you gave, but ALSO wasn't accepted AT ALL by LdB when i presented a more full answer.

Which makes you moronic bitching complaint:

preferring instead to argue and insut

DOUBLEY ironic.

@Michael Kelsey please ignore WOW, I think most get the way he is. He always has to have the last post that is a win in his mind it's like a kid.

If you are here I would much prefer your knowledge put to more useful questions if I may.

I have a couple I would like you to look at

1.) On the exoplanet details I assume the spectra are from when the planet moves in front of it's sun. So am I right in thinking what we are seeing is the dark side of the planet and the sun coming thru a thin atmosphere circle around the dark sphere.

2.) The question that leads to is that spectra really going to be indicative of the near planet side. I have been backward and forward on this one.

The questions are important because there are some reflective moon spectra which you were correct about. However that problem manifests itself.

@LdB #45: In fact, we can (potentially!) get exoplanet spectra both ways. For a transiting exoplanet you can look at the spectrum during a transit and 90-degrees off transit. The difference plot will give you information about the planet's night side (not just atmospheric transmission, but also IR emission).

You can also do spectra just before and after the opposition transit (with the planet moving behind the star), and pick up day-side reflection spectra.

For directly imaged exoplanets, it may be possible to use a fiber mask to isolate a spectrum from the planet image separate from the star.

By Michael Kelsey (not verified) on 29 Sep 2015 #permalink

Cheers for that those are the sort of general details that don't get put in papers because those who work in the area learn that stuff and it doesn't get talked about.

So the emission spectrum will have 4 basic components?

1.) Incident sunlight spectrum
2.) Black body thermal distribution emission
3.) Rayleigh scattering in the atmosphere of 1 & 2
4.) Bulk attenuation by the atmosphere of 1 & 2

And that gives them a night and day emission and even at angles if they so desire in their model that they keep talking about.

Oh look, exactly the same answer as I gave!

And did you ever read the paper you gave a link to? It gave the same answer as I did, along with a very similar explanation of why as I did.

AND gave the answer "Yes, it IS Blue".

Yup, the problem wasn't the answer, was it, retard, it was who gave it.

Your still the only one who believes that !

Oh and there isn't a color only a spectra :-)

That doesn't mean anything. And counterindicated if your demand I be ignored was obeyed, since it means that there is no method by which a valid counterview could be attained with no evidence.

Where in post 18 do you see anything different than what you gratefully accepted from Michael?

EVERY SINGLE STATEMENT in the list in post 47 is in post 18, yet your response was ignorance and bile to 18 and knob gobbling gratitude in post 47.

Where, precisely, did you get your list in there, by the way? where did Michael denote those points?