APOD: Aurora Over Alaska (2010 Oct 06)

[rstevenson]

In case anyone else doesn't know what "forbidden" means in relation to emissions, there is a page about the Forbidden mechanism at Wikipedia. I'm not qualified to judge the content of that page, so maybe one of youse experts can give it a read.

Rob

[Chris Peterson]

I still find it very, very improbable that blue-green light would come out as yellow to yellow-green.

In the first place, while the dominant emission is probably 500.7nm, it is certainly not the only emission. So the apparent color will almost certainly be something other than green. In the second place, it is very possible for a DSLR to produce an unexpected color in a case like this. The sensor passbands are very leaky- that is, even the red channel has signal from a 500.7nm source. Then, internal color mapping tools designed around models of human vision are used to perform color conversion (something also done in RAW file converters). Between these two things, I'd say just about any apparent color might be seen in an aurora image, and two cameras from different manufacturers will probably produce significantly different results.

I agree with you that color images have to be judged with care. That is why I always try to see as many color images as possible of the same object, or of similar objects.

And if the images contain narrow band emission sources, the more images you look at, the more confusing the situation is likely to seem!

[Ann]

Chris wrote:

And if the images contain narrow band emission sources, the more images you look at, the more confusing the situation is likely to seem!

Really? I think that narrowband images are often relaively easy to read, as long as you know what filters they have used and what colors they use to "describe" they have "measured. Take this Hubble image of the Ring Nebula:

Click to view full size image

The simple fact that the nebula is shown to be both very blue in the middle and also yellow near the bright "edge" demonstrates that this must be a narrowband image. The blue center clearly corresponds to oxygen emission, which has been shown as blue in this image. The pale green parts of the nebula must show relatively pure hydrogen emission, which has been shown as green, even though the emission in itself is red. The broad yellow part of the nebula must correspond to a mixture of hydrogen emission and some other emission which I can't remember, but I think it it sulphur or possibly nitrogen. This emission is almost identical in color to hydrogen emission, but in this image only the sulphur (or nitrogen) emission is shown as red. The red "border" shows where there is no longer any hydrogen emission to speak of, but there is still clearly sulphur (or nitrogen) emission.

There! Was that so hard? Admittedly is is so much harder - indeed, sometimes impossible - to "read" the Hubble narrowband images of planetary nebulae, since the "creative coloring" there is beyond belief. It's all in beyonderland, Beyond!

Ann

[Chris Peterson]

Really? I think that narrowband images are often relaively easy to read, as long as you know what filters they have used and what colors they use to "describe" they have "measured. Take this Hubble image of the Ring Nebula...

You misunderstood me, or I'm misunderstanding you. Narrow band images made through narrow band filters are one thing, and I agree they are easy to understand. I was talking about conventional RGB "natural color" images made of objects with strong narrow band emissions. These tend to look very different from image to image, because of all the reasons previously stated.

[zloq]

See my earlier post. The 500.7 [OIII] line is strong in nebulae, while the [OI] line is strong in an aurora and is even a nuisance presence in skyglow. I doubt 500.7 has been detected in an aurora because it requires low density and high electron temperature.

Ann apparently used crude dslr imagery on the web not only to deduce that the color was different from [OIII] in nebulae, but she gave an estimate of 560nm, which is only 2.3nm away from the [OI] auroral line at 557.7.

Frank

[bystander]

The simple fact that the nebula is shown to be both very blue in the middle and also yellow near the bright "edge" demonstrates that this must be a narrowband image. The blue center clearly corresponds to oxygen emission, which has been shown as blue in this image. The pale green parts of the nebula must show relatively pure hydrogen emission, which has been shown as green, even though the emission in itself is red. The broad yellow part of the nebula must correspond to a mixture of hydrogen emission and some other emission which I can't remember, but I think it it sulphur or possibly nitrogen. This emission is almost identical in color to hydrogen emission, but in this image only the sulphur (or nitrogen) emission is shown as red. The red "border" shows where there is no longer any hydrogen emission to speak of, but there is still clearly sulphur (or nitrogen) emission.

If you are really interested, you could just look it up and eliminate the guess work:
Red: F658N ([N II]), Green: F501N ([O III]), Blue: F469N (He II)

http://hubblesite.org/newscenter/archiv ... fastfacts/

At least you got the nitrogen right.

[Chris Peterson]

See my earlier post. The 500.7 [OIII] line is strong in nebulae, while the [OI] line is strong in an aurora and is even a nuisance presence in skyglow. I doubt 500.7 has been detected in an aurora because it requires low density and high electron temperature.

Ann apparently used crude dslr imagery on the web not only to deduce that the color was different from [OIII] in nebulae, but she gave an estimate of 560nm, which is only 2.3nm away from the [OI] auroral line at 557.7.

I agree that the green is more likely from the 557.7nm line than from the 500.7nm line. However, the 500.7nm line is definitely observed in auroras.

BTW, the point of my original post on the matter wasn't based on the specific line or lines involved. The point was that this image shows interference rings because there is a strong monochromatic light source present. Whether this is 557.7nm or 500.7nm doesn't make any difference.

[Ann]

I still find it very, very improbable that blue-green light would come out as yellow to yellow-green.

Ann

Oxygen has many emission lines, including multiple "forbidden" ones. The ones that appear depend on the temperature and density of the plasma. Nebulae are hot and low density, so they emit a lot of [OIII] at 500.7, which is very green. An aurora is much cooler and denser, and its main O emission is [OI] at 557.7nm - more toward the yellow. So - yes - the colors are different because the plasma is very different.

The [OI] line is also present in nebulae, but it's harder to measure since the same line is emitted by our atmosphere, whereas [OIII] is not. In order to measure [OI] in astronomical objects, they look for fast moving ones that are doppler shifted from the terrestrial emission. See Keenan et al., Forbidden Lines of [OI] in the High-Resolution Optical Spectra of Planetary Nebulae, Pub. Astro. Soc. Pac., 107, 1995.

Thank you so much, zlog! That is exactly what I thought. I have looked at many, many images of aurorae, and I have seen several which have been "all green", but not a single one has ever looked "all blue". OIII emission is close enough to the blue part of the spectrum that if monocolored OIII aurorae were common, then there would undoubtedly exist photographs of aurorae showing aurorae as "all blue". But I have never seen a picture which made an aurora look like that.

Click to view full size image

These are the spectral lines of oxygen, according to wikipedia. The OIII line is probably the brightest of a pair of lines that appear green, yet is not so far from the "blue" part of the spectrum.

Yes, but look at the two brightest green lines. I would assume that one of them, probably the brightest one, corresponds to a wavelength of 557.7nm. and in my opinion, there really are so many aurorae that photograph just like that and appear to be just that color!

So in my opinion, the most likely dominant emission of the aurorae over Alaska would indeed be OI emission at 557.7nm.

Ann

[Ann]

The simple fact that the nebula is shown to be both very blue in the middle and also yellow near the bright "edge" demonstrates that this must be a narrowband image. The blue center clearly corresponds to oxygen emission, which has been shown as blue in this image. The pale green parts of the nebula must show relatively pure hydrogen emission, which has been shown as green, even though the emission in itself is red. The broad yellow part of the nebula must correspond to a mixture of hydrogen emission and some other emission which I can't remember, but I think it it sulphur or possibly nitrogen. This emission is almost identical in color to hydrogen emission, but in this image only the sulphur (or nitrogen) emission is shown as red. The red "border" shows where there is no longer any hydrogen emission to speak of, but there is still clearly sulphur (or nitrogen) emission.

If you are really interested, you could just look it up and eliminate the guess work:
Red: F658N ([N II]), Green: F501N ([O III]), Blue: F469N (He II)

http://hubblesite.org/newscenter/archiv ... fastfacts/

At least you got the nitrogen right.

Well, frankly, no, bystander. I'm not too interested in planetary nebulae. I find their colors non-beautiful even in a best-case scenario, and the utter, total confusion of the colors used in most Hubble pictures of planetaries have made me almost dislike these objects.

Indeed, I was wrong about the Ring nebula. Even so I was right, because the bluest part corresponded to the highest ionization and highest temperature, while the reddest part corresponded to the lowest ionization and lowest temperature. What I find most interesting here is that the Hubble has not used a filter for hydrogen. Is there no hydrogen in this nebula? Obviously there are some nebulae which are devoid of hydrogen.

But all in all, I don't care too much about these critters. Thanks anyway, bystander, for setting me straight.

Ann

[neufer]

.................................................................................
The blue center clearly corresponds to oxygen emission, which has been shown as blue in this image.

The pale green parts of the nebula must show relatively pure hydrogen emission, which has been shown as green, even though the emission in itself is red.

The broad yellow part of the nebula must correspond to a mixture of hydrogen emission and some other emission which I can't remember, but I think it is sulfur or possibly nitrogen.

This emission is almost identical in color to hydrogen emission, but in this image only the sulphur (or nitrogen) emission is shown as red. The red "border" shows where there is no longer any hydrogen emission to speak of, but there is still clearly sulphur (or nitrogen) emission.
.................................................................................
There! Was that so hard? Admittedly it is so much harder - indeed, sometimes impossible - to "read" the Hubble narrowband images of planetary nebulae,

Apparently so (without a program) :

NASA's Hubble Space Telescope has captured the sharpest view yet of the most famous of all planetary nebulae: the Ring Nebula (M57). In this October 1998 image, the telescope has looked down a barrel of gas cast off by a dying star thousands of years ago. This photo reveals elongated dark clumps of material embedded in the gas at the edge of the nebula; the dying central star floating in a blue haze of hot gas. The nebula is about a light-year in diameter and is located some 2000 light-years from Earth in the direction of the constellation Lyra.

The colors are approximately true colors. The color image was assembled from three black-and-white photos taken through different color filters with the Hubble telescope's Wide Field Planetary Camera 2.
.................................................................................
Blue isolates emission from very hot helium, which is located primarily close to the hot central star.

Green represents ionized oxygen [as in green aurora], which is located farther from the star.

Red shows ionized nitrogen [as in red aurora], which is radiated from the coolest gas, located farthest from the star.
.................................................................................
The gradations of color illustrate how the gas glows because it is bathed in ultraviolet radiation from the remnant central star, whose surface temperature is a white-hot 120,000 degrees Celsius (216,000 degrees Fahrenheit).

[Chris Peterson]

These are the spectral lines of oxygen, according to wikipedia. The OIII line is probably the brightest of a pair of lines that appear green, yet is not so far from the "blue" part of the spectrum.

I don't see how you can draw any comparisons between the aurora image and the narrow band HST image of the Ring. The latter clearly defines specific emissions, which are then mapped to red, green, and blue. The aurora image irretrievably convolves several emission lines and a background continuum. It is somewhere between very difficult (with the raw data) and impossible to determine from the DSLR image what the actual emission lines are.

[neufer]

Indeed, I was wrong about the Ring nebula. Even so I was right, because the bluest part corresponded to the highest ionization and highest temperature, while the reddest part corresponded to the lowest ionization and lowest temperature. What I find most interesting here is that the Hubble has not used a filter for hydrogen. Is there no hydrogen in this nebula? Obviously there are some nebulae which are devoid of hydrogen.

Planetary nebula are stars that have evaporated all of their meager supply of 'unburned' hydrogen in earlier stellar winds.

[zloq]

I agree that the green is more likely from the 557.7nm line than from the 500.7nm line. However, the 500.7nm line is definitely observed in auroras.

Do you have a reference for 500.7 being detected in an aurora - particularly at a level that it is visible to the eye?

[Ann]

Anyway, let me point out one thing. I never questioned Chris' claim that the concentric rings in the image were caused by monochromatic light. I just said that this monochromatic light couldn't be OIII emission. Chris refused to admit that I could be right about this. When he himself was shown to be wrong, he just shrugged and said that it doesn't matter that he was wrong, because he was right in the only way that matters.

So I will assume Chris' way of arguing and say that it doesn't matter that I was wrong about the Ring Nebula, because I was right in the only way that matters.

Ann

[Chris Peterson]

Do you have a reference for 500.7 being detected in an aurora - particularly at a level that it is visible to the eye?

I have some images from an array of near-Arctic allsky cameras, utilizing different filters. I designed the cameras, but am not the researcher running them. I don't know what has been published, but I'll see if I can find out, or get permission to post an image.

The cameras use unintensified video, so anything they show is easily bright enough to be visible to the eye.

[Ann]

These are the spectral lines of oxygen, according to wikipedia. The OIII line is probably the brightest of a pair of lines that appear green, yet is not so far from the "blue" part of the spectrum.

I don't see how you can draw any comparisons between the aurora image and the narrow band HST image of the Ring. The latter clearly defines specific emissions, which are then mapped to red, green, and blue. The aurora image irretrievably convolves several emission lines and a background continuum. It is somewhere between very difficult (with the raw data) and impossible to determine from the DSLR image what the actual emission lines are.

No, Chris, I'm not comparing the color of the aurora in yesterday's APOD with the color of the Ring Nebula in the Hubble image. Because, indeed, I don't compare the color of broadband images with the color of narrowband images.

I'm comparing the color of the 557.7nm line in the wikipedia image with the color of a lot of broadband images of aurorae that I have seen.

Ann

[Chris Peterson]

Anyway, let me point out one thing. I never questioned Chris' claim that the concentric rings in the image were caused by monochromatic light. I just said that this monochromatic light couldn't be OIII emission. Chris refused to admit that I could be right about this. When he himself was shown to be wrong, he just shrugged and said that it doesn't matter that he was wrong, because he was right in the only way that matters.

Sorry if I gave that impression. I don't think I said you were wrong about what emission lines you think are involved. I was only questioning how effectively you can identify narrow emission lines using a DSLR or other one-shot color camera.

It is possible to look at the physics of auroras, and make some reasonable assumptions about what is causing the colors. But I don't think you can look at aurora images and at narrow band (or even RGB) astroimages and use perceptual color comparisons to determine the spectral lines found in this aurora image.

[Ann]

Chris wrote:

Sorry if I gave that impression. I don't think I said you were wrong about what emission lines you think are involved. I was only questioning how effectively you can identify narrow emission lines using a DSLR or other one-shot color camera.

Well, I think I can do it, and Frank - zloq - says that I was able to do it.

Ann

[neufer]

These are the spectral lines of oxygen, according to wikipedia. The OIII line is probably the brightest of a pair of lines that appear green, yet is not so far from the "blue" part of the spectrum.

I don't see how you can draw any comparisons between the aurora image and the narrow band HST image of the Ring. The latter clearly defines specific emissions, which are then mapped to red, green, and blue. The aurora image irretrievably convolves several emission lines and a background continuum. It is somewhere between very difficult (with the raw data) and impossible to determine from the DSLR image what the actual emission lines are.

Ann was actually pretty close on this one. Here is the Earth imitating the Ring Nebula:

See http://earthobservatory.nasa.gov/Features/ISSAurora/

---

See http://earthobservatory.nasa.gov/Features/ISSAurora/

---

[Ann]

Let me explain once again how I could be so sure that this aurora could not be monochromatic and dominated by OIII emission at the same time.

I make a note of the color of any object I see, certainly of every brightly colored object. I don't remember the color of everything I see, but I remember a lot, and I always know if a particular object has a color that is unusual or unexpected for this particular object.

I have never seen an aurora in real life, but you can't spend a lot of time looking at astronomical objects without seeing a lot of aurora images, too. I always note the color of the aurorae. And since blue is far and away my favorite color, I am always actively looking for blue things. I am always asking myself, could this be a blue object? When it comes to aurorae, I have never been able to say a definite "yes" to that question. At best, multicolored aurorae have looked a bit blue in places, but I have never been able to trust that color. On the other hand, many aurorae have looked green. Quite often they have been a sort of "in-between-green", neither yellow-green nor blue-green. Several times, however, they have been strongly yellow-green, so much so that their yellow-green color has positively "jumped out" at me. Their intense yellow-green color has made a strong impression on me.

But they have never looked truly blue. Never.

I am well aware that you can never trust the color balance of an individual image. The color balance can always be off. But I know, I trust, that if a particular object is often seen to be a monocolored "in-between-green", and sometimes it is an intensely monocolored yellow-green, then I can be sure that this particular object can be a monocolored green, and its shade of green is can be either "in between" or yellowish, but i can never be bluish. Because if the "true" color of the object was really blue-green, then I would sooner or later see a picture of this object where the color balance was sufficiently off for the blue-green color to look blue. In other words, if monocolored blue-green aurorae were the least bit common, then I would have seen a picture of what looked like a monocolored blue auroa. But I have never seen that, never. And I know that I haven't, because if I had, I would definitely remember it. Therefore, I would go so far as to say that monocolored aurorae dominated by strongly blue-green OIII emission either don't exist at all, or at least they are extremely uncommon.

Chris said that the aurora over Alaska was monocolored, since the camera or film reacted to its monochromacity by producing concentric rings. I have no reason whatsoever to doubt Chris' assessment that the aurora is monochromatic. I have seen many images of green aurorae that have definitely looked monochromatic to me.

But I strongly doubt the existence of monochromatic OIII aurorae, like I said. By contrast, I know that many planetary nebulae are dominated by OIII emission. I also know that a Google search will turn up a planetary nebula that looks more blue than green. I'm not talking about those popular creatively-colored narrowband images of planetaries, but probable broadband ones. This is a typical broadband image of a planetary dominated by OIII emission, where the image makes the planetary look blue:



I know that the color balance is off here. I can see that it is. I know that it is, because astroimagers who are serious about their colors almost always make their OIII nebulae look greenish. But an image like this one confirms my suspicion that the "true" color of an OIII nebula is close enough to the blue part of the spectrum for a few astroimages to make the nebula look blue. The more blue-looking planetaries I can find, the more certain I can be that OIII emission is indeed blue-green in color. I remember very well that I have seen other broadband astroimages of blue-looking planetaries than the one I posted above.

But I have, as I said, never seen an aurora look like that. Never. Therefore I'm sure that aurorae can't be dominated by OIII emission. I strongly, strongly doubt that big, bright, monochromatic OIII aurorae even exist.

And if someone tells me that a bright yellow-looking aurorae is both monochromatic and dominated by OIII emission, I'm not going to believe them. If an aurora looks yellow, it must either be a mixture of two different kinds of emission, or else the color balance of the picture must be severely off. Indeed, the color balance of the picture seemed definitely too red. So the aurora may well be green instead of yellow. I know that monocolored yellow-green auroae definitely exist, and they are even pretty common. So this could be a monocolored yellow-green aurora. If so, its dominant color should correspond to a wavelength of about 560nm, because the yellow-green aurorae I have seen appear to "cluster" around a wavelength close to 560nm, based on the average appearance of them.

Elementary, my dear Watson.

Ann

[Chris Peterson]

Chris said that the aurora over Alaska was monocolored, since the camera or film reacted to its monochromacity by producing concentric rings.

Actually, that's not what I said. To be clear, what I said was that there was a strong monochromatic source present in the scene, and this is why we could see interference rings. As in any aurora, we can expect several emission wavelengths to be present. In addition, this scene is also lit by a continuum source. All of this makes the perception of color complex.

Color is not a physical thing, but a physiological one, and the aurora is very obviously not monocolored.

[Ann]

Well, wow. I did another auroa search, and for the first time ever I found a blue-looking broadband aurora image:



The color balance is clearly off. Nevertheless, I wouldn't swear that this aurora isn't dominated by OIII emission, based on the appearance of it. And the existence of this picture lends some credence to the idea that aurorae dominated by OIII emission may exist. But I am still very skeptical.

Ann

[Chris Peterson]

Well, wow. I did another auroa search, and for the first time ever I found a blue-looking broadband aurora image...

I've seen several high latitude auroras, and several low latitude auroras. All of the high latitude events looked very similar to what this image shows- blues with a bit of purple. I never saw much green. All of the low latitude events were mostly deep red, with green below the red. No blue or purple. The differences make sense considering the very different energies involved in low and high latitude auroras.

[Guest]

And the existence of this picture lends some credence to the idea that aurorae dominated by OIII emission may exist. But I am still very skeptical.

Ann

The aurora environment is very different from a gaseous or planetary nebula, so there is no reason to expect similar emission lines. The main lines people image in nebulae are Ha, [SII], [OIII], and [NII]. The latter three ines require atoms to be highly ionized by UV, then collisionally excited by electrons, and then remain in that state for 10's to thousands of seconds in order to emit the particular photon through a "forbidden" transition. This requires a combination of strong ionization source, high temperature electron gas, and extremely low density. Although oxygen and nitrogen are abundant in aurorae, the other conditions of a strong ionization source, high temperature electron gas (10000K), and extremely low density (to avoid collisions that would de-excite the ion and stop it from emitting the desired wavelength) do not exist. A tropical fish contains both O and N, but I would not conclude that its bright colors are due to forbidden lines of highly ionized atoms. The [OI] line requires less ionization and has a shorter lifetime - and the presence of the O ion is aided by a chemical interaction with N2 molecules - so it is a strong line in aurorae and skyglow.

There are many emission lines across the spectrum that contribute to the colors in an aurora - so I wouldn't call them monochromatic - although the color is dominated by a handful of lines. I could not find a good paper summarizing all the lines, but there are many web pages showing the main ones. There is a book called, "Physics of the aurora and airglow" that is viewable on google, and it contains a large atlas of auroral lines starting at around page 150. There are many of them, and they are very different from what you would find dominant in a nebula.

Note also that colors in an aurora are due to a region only 10's to 100's of km deep - while in a nebula it is 10's to 100's of light years deep. At the required low density to allow [OIII] and [NII] lines to appear, you need a large depth of material to create enough glow to be visible.

So - you noticed that the color in the APOD aurora image did not look like 500nm and instead seemed closer to 560nm - and that was an accurate assessment.