Clearing the air (or, Mea Culpa Part 1)

By Phil Plait | December 21, 2010 7:02 am

Like most of you, I’m human. I try to be as accurate as possible when I write, but sometimes I make mistakes. A lot of these errors are small and I just fix ’em. Some of them are bigger, and I generally strike them through and correct them up front — leaving them public keeps me honest. Also, part of science is learning from your mistakes. If we don’t, we get ossified and dogmatic, and that’s the very antithesis of science!

Also, sometimes, these mistakes deserve more airtime. They deserve their own post, and it so happens I have a couple on which I’d like to elaborate. I want to clear the air, so to speak, and what better way to start than to talk about clear air?

In the past, when giving talks (as well as in my first book) I say that the Earth’s air is very transparent. In fact, based on the memory of a paper I read in grad school and which I can no longer find, I’ve said specifically that 98% of the visible light that enters our atmosphere from space will make it to the ground (barring clouds and so on). This is usually in response to people asking if they’d see more stars from space due to the lack of air, and my answer has always been "not really", most of the light from even faint stars makes it to your eye.

That turns out not to be correct.

To illustrate this, here’s a graph showing an example of how much light from the Sun actually gets to the ground:


The details of this are complicated, but it illustrates the point. This graph is from the American Society for Testing and Materials, who created it to aid the solar power community. The vertical axis is brightness (specifically, the amount of energy hitting an area every second), and the horizontal axis is wavelength (think of it as color). Visible light has a wavelength of very roughly 300 – 700 nm (nanometers, a billionth of a meter) on that scale. The black line is a model of the Sun’s light, and the blue line is the amount of that which hits an area tilted up to face the Sun (details are at the link above).

According to this, in visible light quite a bit of power from the Sun is lost through the air, roughly 20-25%. This is a pretty clear (haha!) indication that the atmosphere does rob us of some light.

What happens to that light? Well, some of it is simply absorbed by stuff in the air. That light is essentially gone, sucked up by atoms and molecules and turned into heat. Other light gets scattered; imagine a photon of light is like a ball in a pinball game and the atoms in the air are the bumpers, to get the picture. As the light comes down, it bounces off the atoms and goes every which way. This tends to diffuse light a bit, making stars fainter. Also, blue light gets scattered more than red, which is why the sky is blue: those photons from the Sun get scattered all over the sky, so you see blue light coming from every direction. Tadaa! Blue sky.

There’s more, too. Sky and Telescope has a good explanation of all this. One big factor is where in the sky you’re looking. The Earth’s atmosphere is a thin shell above us, and the Earth is a ball underfoot. That means that when you look straight up, you’re looking through less air than you do toward the horizon. Here’s a diagram:


The inner circle is the surface of the Earth, and the outer one represents the top of the atmosphere. The vertical red line is how much air you look through when you look straight up; the horizontal one is when you look toward the horizon. See how much longer the path length is for light coming from a star on the horizon? That’s why, on a dark night, you see more stars overhead than lower down. It’s also why astronomers like to wait for their targets to rise up high before they observe them! [If this looks familiar, it’s similar in principal to what I was describing in a recent post about seeing ring-shaped nebulae in the sky.]

We even have a term for this: air mass. When you look straight up, you’re seeing through one air mass. The lower you look, the higher the air mass, and the worse your object gets.

As you might expect if you’ve gotten this far, I’m simplifying things a lot here! There are tons of details dealing with all manners of rotten air quality, but the point here is that our air does mess up our view, and starlight from space doesn’t all get to your eye. I’ll note, though, that our air really is remarkably transparent; given that it’s many kilometers thick it’s pretty cool that any light gets through at all. As you can see from the top graph, a lot of ultraviolet (shorter wavelength) and infrared (longer wavelength) light never gets to the ground. But then, there’s a really good reason we call the light we can see visible light. It’s no coincidence; it makes sense from evolution that we see best in the kind of light that makes it down here from the Sun.

And of course, all of this is why we launch telescopes into space. No air means more light gets to the telescope (including wavelengths that otherwise never reach the ground, again like IR and UV), as well as losing the annoying twinkling caused by our roiling, moving atmosphere. Twinkling stars are fine for poetry, or for lovers holding hands and stargazing, but it’s a nuisance when you’re trying to see two objects close together that get blurred into one blob!

And finally, what does this mean for seeing stars in space? Well, air removes 25% of the light from a star. In astronomical terms, that’s equivalent to about 0.25 magnitudes. The faintest star you can see from the ground is about magnitude 6 or a bit fainter, so pushing it I’d say you could see stars as faint as magnitude 6.5. That means there are a few thousand more stars that would be visible. Under ideal circumstances, the number of stars you could theoretically see would jump from about 6000 to about 8000. I’m being pretty rough here, but the point is the number of stars would increase, but not hugely. And that would be on the fainter end, making them harder to see anyway. So if you were in space, you might expect to see more stars, but not be overwhelmed by them!

So that’s it. Thanks for indulging me. I’ll always strive for accuracy, but when things go wrong, I’ll post another "mea culpa"… and in fact, another one will be cometing — I’m sorry, coming pretty soon. Hint hint.

My sincere thanks to amateur astronomer George Cooper, who sent me emails about this and other interesting topics over the years.

CATEGORIZED UNDER: About this blog, Astronomy, Science
MORE ABOUT: air, Mea Culpa, scattering

Comments (52)

  1. Blizzzzzaaaarrg!

    Sun, I am disappoint.

  2. Captn Tommy


    Merry Christmas To you and your’s.

    and thanks for your Blog

    Captn Tommy

  3. Gus Snarp

    Hah! I knew something you didn’t! I did my master’s work on solar energy, so I knew this. I did not, however, know that you didn’t know.

    But really, I was fully expecting after the introduction that you were going to explain the disappearing laser post…

  4. Jeremy

    I was under the impression that astronauts claimed to be able to see quite a bit more stars in space, but upon googling haven’t been able to find anything.

    Astronauts in orbit, if they were on the unlit side of the Earth, and they shut off the lights on the orbiter (or ISS), they would only see 8,000 stars?

  5. Dan Oberste

    BUT!!! The 25% increase would be only for the stars directly overhead. As the angle to the horizon becomes more acute, the increasing airmass would shield even more. PLUS, as you increase your altitude the horizon falls, revealing even more area of sky to view stars in.
    So I say the number of additional stars viewable from space might very well double.
    What say you Phil?

  6. And let’s not forget about the effects of “light pollution”, which, although doesn’t stop any starlight from reaching our eyes, does make it harder to see the dimmer objects.

    Several months ago, I was in a location far from any city lights. There was no Moon out that night, and the sky was so dark, that I could see many more stars (and even the Milky Way) in the sky than I do from home, despite it being the same atmosphere.

  7. Brett from Canada

    @Captn Tommy:


    On no you didn’t!

  8. John K

    Jeremy try searching on ISS northern lights.

  9. Daniel J. Andrews

    I thought in ideal locations you could see stars down to about magnitude 7 or even close to 8? Apparently Clyde Tombaugh could see quite dim stars and this was tested by either Sky and Telescope or Astronomy Today (I think the former). They also had an article at one point about the faintest star observed by someone using an oxygen tank to help supplement their vision (don’t remember the details–this would be from the late 80s).

    And what *did* happen to the laser post and why? Just curious now.

  10. John K


    I’m surprised that you didn’t know of the atmosphere’s absorbing and reflecting characteristics. Also, one’s latitude plays a small role too.

  11. rob

    it’s okay Phil. you can’t be right every time. i am not right every time either.

    why just the other day, i thought i was wrong, but it turns out i was really right.


  12. Aerimus

    I was also expecting a comment regarding the missing laser post. One day….

  13. Gary Ansorge

    Ah, the limitations of the human sense organs. From blue to red(roughly 400 to 750 nm) our visual pigments are sensitive but that’s only a fraction of the total solar output. Interestingly, plants are most effective at using just two ranges of the solar spectra for photosynthesis, from about 400 nm to 475 and from 600 to 700 nm. This is the range for chlorophyl A, which provides the energy for food production in the plant. All the rest of the spectrum is ignored/rejected(ok, there IS a small amount of activity elsewhere, but that’s another chlorophyl and is apparently not required to maintain plant metabolism and growth). Green, of course, is totally rejected.

    Dr. Gerard K. O’Neille ( writing about solar power sats)used the figure of 1350 watts/m^2 as the total solar intensity in space at earth orbit and 400 watts/m^2 for the solar intensity at high noon, on the equator, on the earths surface. I just automatically assumed that was mostly in the visible spectrum.

    Thanks for shining a light in that dark corner of my ignorance,,,

    Gary 7

  14. Tim G

    Not much around 1400 nm, 1900nm and 2600 nm is making it to the ground. A large log log graph would look nice.

    (5/2) * log10(3/4) is about -0.31. So a loss of 25% of light would mean a star dimmer by 0.31 magnitudes.

  15. Alex

    What sort of effect could light pollution play? In a big city I bet it makes a huge difference to be 300km above all those street lamps.

  16. Navneeth

    What’s the red part of the graph, Phil? (I’m guessing sunlight received at an oblique angle.)

    And what’s this thing about a “missing laser post” anyway?

  17. Light Pollution

    I suggest you do another followup, Phil, to discuss light pollution. As other commenters have pointed out, that’s a major effect that you seem to be neglecting. Most of us can’t get to the perfectly black place your analysis seems to imply. The signal to noise ratio of our detection system, even when it’s our eyes, is strongly dependent on the background light level scattering off the atmosphere. Try looking for stars in Manhattan. You might see Sirius on a good night. Now go to the Mojave on the same night and you’ll see thousands of stars.

    I was recently amazed to easily see not only the Milky Way, but structure within it out the window of a car driving through the high Sierra. There were plenty of little bright lights from buttons and dials in the car, too, but without the scattered noise in the atmosphere to which I’m accustomed the sky was shockingly full of stars.

  18. Joe Nemanick

    Dr. Plait, I believe that you should not use the graph of energy (or power) vs. wavelength for your graph, but rather flux or photon count vs. wavelength. An energy graph overstates the losses to visibility in the blue part of the spectrum from Rayleigh and Raman scattering. Our eyes detect photons, not their energy, and are more sensitive to photons from the red and green regions of the spectrum anyway.

    Also, it wasn’t clear over what wavelengths you were integrating over to estimate the losses at ~25%. The large amount of IR absorption from water vibrations and diatomics should not impact visibility to the human eye, nor should losses above about 350 nm from scattering and UV absorptions. A better graph would leave off data that is outside of human perception.

  19. @Alex (No. 15) light pollution lowers the contrast between the sky and the stars. We actually don’t see stars as pinpoints of light but rather as specks “smeared” to the resolving power of our eye. If the surface brightness of the sky is higher you won’t see dimmer stars.

  20. Michel

    You are saying this so we will buy bigger scopes!
    Is it Orion or Celestron that makes you say this?
    Sooooo obvious just a few days before cristmas.

    OT: wouldn´t it be great if all the amateur astronomers would put few dollars in the pot so we can launch (hitch a ride) our own hobbyscope up there? Make it a international competition for universities/tech schools? Like Masterchef/Hellskitchen but better.

  21. rob

    Phil had a post about a 1 watt hand held laser. i suspect he took it down cuz a 1 watt laser is NOT a toy. it can blind you at 149 meters with 0.25 sec exposure. if you used it as a cat toy you would burn lines in your floor and perhaps in your cat.

  22. Navneeth

    21. rob

    Phil had a post about a 1 watt hand held laser. i suspect he took it down cuz a 1 watt laser is NOT a toy. it can blind you at 149 meters with 0.25 sec exposure. if you used it as a cat toy you would burn lines in your floor and perhaps in your cat.

    Ah, thanks!

    Do you know there is a whole forum for laser pointers out there? 😮

  23. Michel

    My green laser pointer has a build in safety lock which needs a key.
    It´s cheap ($35), Chinese (, but is the only one I know off that has that feature.
    And that´s how all those pointer devices should be. No one can accidently be “funny” with it.
    Sorry about the spammyness, but a safer laser is always a better laser.

  24. MadScientist

    Make that ~400 to ~700nm for the visible spectrum. 300nm is UV, 400nm should be a beautiful purple color – well, for those people who can see it. And objects are preferably viewed higher off the horizon because (a) turbulence is less and (b) distortion due to refraction in the atmosphere is less (flattening, apparent location).

  25. chris j.

    wasn’t the blue sky (and light pollution) a dead giveaway that the air blocks a lot of light, even without the fancy calculations? and hasn’t the concern about avoiding near-horizon objects been more about the seeing, and not so much about atmospheric extinction?

    i suppose, however, that it could be difficult to grasp how much light our air can block, unless you’ve lived in los angeles.

  26. MaDeR

    “Do you know there is a whole forum for laser pointers out there”
    There are forums for anything. ANYTHING.

    Let it sink into you.

  27. lemuet

    What? PHIL IS HUMAN?!

  28. I’ve noticed that one quite often finds errors on the internet. If we don’t hold back from correcting them, one day they’ll all be gone.

  29. Michel

    @27. lemuet
    Nah, he´s a forum.

  30. Just wondering, if prior to Noahs flood 1n 4400BC, if the canopy of water that surrounded the earth would have magnified starlight like the lens of a telescope and people like Adam and Eve could have seen many more stars…? 😉 :)

    (Note: the smileys if you think I’m serious)

  31. Joseph G

    @#1 Blizzzzzzz:

    @#9 Daniel: So if I suck pure, pressurized oxygen, my eyes will get more sensitive, temporarily?
    Until the oxygen toxicity makes my brain bleed, anyway.

  32. Joseph G

    What are those big dips in the red and blue curves? Are those absorption lines for certain gases in the atmosphere?

    @#21 rob: I had no idea a 1 watt laser was so powerful. I guess it’s the inverse-square law at work? Or lack thereof?

    @#26 mader: Corollary to rule 34: If it exists, there is a forum for it.
    Incidentally, I’ll never be able to pet a horse again.

  33. Gus Snarp

    @Navneeth – Since this is used by solar power folks, I think the red is the radiation on a panel parallel to the earth while the blue is a panel tilted to the optimum for that latitude.

    @Joseph G – Yeah, basically. The dips can be absorption, but also reflection.

  34. Sven

    well, you told us to be skeptic anyway so… it’s ok

  35. Joseph G

    Has anyone mentioned the SOFIA yet? It’s a very cool project (literally as well as figuratively).
    Of course, it’s observing in the infrared, which is blocked by water vapor in the atmosphere to a much greater extent than visible light, so it’s not really the same issue. I’m sure the low temperatures at altitude help, too.

  36. Bioth

    Nice to see that the graph uses SI units, and not them stupid janskys or even worse: cgs units!

  37. CB

    I suggest you do another followup, Phil, to discuss light pollution. As other commenters have pointed out, that’s a major effect that you seem to be neglecting. [snip] Try looking for stars in Manhattan. You might see Sirius on a good night. Now go to the Mojave on the same night and you’ll see thousands of stars.

    What do we need Phil say? What is there to say, that you didn’t just say yourself? With heavy light pollution, you’ll see few stars if any. Move to where the sky is dark, and you’ll see many more. Everyone already knows this. Heck, I was on the Strip in Las Vegas, and the only thing you could see in the sky was Jupiter (which I thought was pretty impressive of ol’ Jupes).

  38. mike burkhart

    Nobodys perfect I have made plenty of mistakes (like my spelling).Your right that science is not dogmatic but I have found that some scientists do cling to pet theroys and refuse to give them up even when proven worng ,(like Dr Fred Hoyel who utill the day of his death still clung to his stedy state theroy even when most Astronomers discarded it) Also science maybe slow to accept new theroys because they challenge covental views.Not of corse that science should accept every idea,but only thoses that have proof and are well reserched..

  39. aronne

    The spectral features (the “dips”) are indeed molecular absorption. Most of the wide features are absorption by water vapor. The small wiggle at 2000 nm is carbon dioxide, and I think the narrow line at 750 nm is oxygen. The broad decrease (primarily less than 900 nm) is rayleigh scattering. The scattering is not significant above 1000 nm which is why all 3 lines overlap at those wavelengths when you are not looking at a molecular absorption feature.

    As for the red and blue lines – I stared at this for a while but I think I finally figured out. The red line is the direct solar irradiance after absorption by 1.5 airmasses, but the blue line is the direct solar plus the scattered solar in the whole hemisphere seen by the solar cell. Basically, if you did a spherical integration over the hemisphere except where the sun is (e.g., collect all the blue sky), you would get the difference between the blue and the red curves. The difference between the blue and black curves in the visible is then light that is scattered back into space.

    So for the problem considered, the red line would be the one to consider for starlight. The blue curve is only relevant to a solar cell that collects light over the whole hemisphere.

  40. Mike Saunders

    Still not entirely right, the sun outputs energy in a much larger spread of wavelengths than introduced by that graph. And the atmosphere positively absorbs most of it, there’s just a giant notch in the area of visible light that it lets through. Probably why we can see light at that frequency and not others, it just doesn’t get through.

  41. Pete Jackson

    When I used to do UBV (ultraviolet, blue, visual) photometry at 7000 feet elevation in Chile back in the 1970s, the effects of the atmosphere were measured by repeating observations on standard stars every half hour or so. The drop in signal as a star moved closer to the horizon was plotted against the zenith distance (degrees in the sky from the zenith to the star). Reading the graph would give you the extinction (light loss due to absorption and scattering) per air mass, a quantity normalized to unity at the zenith. Typically, extinction at the zenith at that elevation was about 15% in V, 25% in B and about 55% in U. At sea level, extinction will be about 50% more (22.5, 37.5, and 77.5).

    Dust and volcanic haze will increase the extinction at all levels, but in a fairly ‘gray’ (wavelength independent) way.

  42. Ze Kraggash

    Phil said, “But then, there’s a really good reason we call the light we can see visible light. It’s no coincidence; it makes sense from evolution that we see best in the kind of light that makes it down here from the Sun.”

    This is actually a triple coincidence:
    1. Visible light is the peak of the Sun’s radiation curve. This would not be the case if it were much hotter or cooler.
    2. The atmosphere is much more transparent to visible light than longer or shorter wavelengths.
    3. Visible light has enough energy per photon to induce chemical effects (i. e., vision) but not so much as to damage biological material.

  43. Whatever small gain you would get by going above the atmosphere would probably be canceled out by the light loss in your spacesuit’s helmet’s faceplate’s transmittance.

  44. Messier Tidy Upper

    Hey, we’re all fallible humans and we all have our moments – some of us* just have more moments than others. 😉

    Nothing too unusual or to be too ashamed of there methinks. Often we learn more through our mistakes than what we get right.

    @38. mike burkhart : “Nobodys perfect I have made plenty of mistakes like my spelling.”

    Agreed there. As for typos & spelling errors .. I could swear that my #@!&^$^$#@@! computer adds them to my posts in just that final second before my editing time runs out! :-(

    .. & when I see them I usually do! 😉


    * & mea culpa myself there.

  45. Messier Tidy Upper

    @37. CB :

    What do we need Phil say? What is there to say, that you didn’t just say yourself? With heavy light pollution, you’ll see few stars if any. Move to where the sky is dark, and you’ll see many more. Everyone already knows this. Heck, I was on the Strip in Las Vegas, and the only thing you could see in the sky was Jupiter (which I thought was pretty impressive of ol’ Jupes).</i.

    The celestial object that most impressed me versus overwhelming light pollution was comet McNaught (2007) seen from the Adelaide oval in the centre of my (admittedly small by international standards) city with all the light towers on in evening twilight around sunset but whilst still quite light after a day-night cricket game. Now *that* was impressive! 8)

  46. MadScientist

    @Joseph#32: Not necessarily – look at the spectrum taken from space (black). Many of the lines are features of the sun’s own light (see the wikipedia article on Fraunhofer).

    Now there’s a huge bite taken out below 400nm; that is partially due to scattering and closer to 300nm it is due primarily to oxygen and ozone (but there is also significant absorption by other trace gases); below ~280nm water vapor also absorbs UV very well and the atmosphere is pretty much opaque.

    On the longer wavelengths it should be obvious that many of the features are due to absorption by the atmosphere. Astronomers with instruments on the ground are forced to study the near infrared region only in those parts where the atmosphere is transparent; there are no such restrictions in space.

  47. Jeffersonian

    Since we stick our telescopes at 13,600′ on Mauna Kea and 14,000′ on Mt Evans, I always figured it goes without saying. I’ve certainly seen more stars above 13k, personally, in particular the Milky Way is more obvious.

  48. Bruce

    I can’t wait to read your apology when you admit that global warming is complete BS.

  49. réalta fuar

    Hard to know which is more astonishing, that a professional astronomer didn’t know the term extinction coefficient or that one got away with saying something so trivially dumb for so long. I’d fail a first year for not knowing this.
    If stars are all of equal brightness and evenly distributed (not a bad first zeroth approximation for the disk of the galaxy), you’d see about 40% more stars from above the atmosphere. And this of course doesn’t account for the much DARKER sky, so I’d say a good first estimate would be closer to twice as many stars visible to the naked eye from space.

  50. Buzz Parsec

    Michel @ 20,

    There was a project by the Independent Space Research Group (originally a group of RPI students) that was building a 50cm telescope to be launched as a get-away special (GAS) cannister on the Shuttle in 1987. They disappeared, and I can find little about them on the web. I think they lost their ride after the Challenger disaster. I found a link to them, but it appears the server is down. It might eventually work.

    There was also a project to build an amateur telescope to be mounted on the ISS (the ISS-AT) , but the project web site is defunct. Maybe it ran into problems after Columbia?

  51. Buzz Parsec

    Bruce @ 48, you mean like evolution and gravity? You are wrong, wishful thinking doesn’t trump science. Nice trolly, here’s a treat.

    Everyone else, sorry.

  52. réalta fuar (#49): Of course I know about extinction, and never said I didn’t. However, as I said in the post, from what I read in grad school, under ideal conditions the extinction for visible light was very near 0 (well, 0.02). That turns out to be incorrect, which is the point of the post. It’s possible I was misremembering the absorption for a narrow line and that got stuck in my head as broadband.

    I’m not sure why you decided to be so snide in your comment; especially since my reasoning was clear in this post — and I’ll note you inferred I didn’t know what an extinction coefficient is, which is simply not true.

    And if you’d fail a first year for a mistake like that… well, I’m glad you were never my professor. Better students learn from mistakes than be punished for them.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


See More

Collapse bottom bar