APOD: Unusual Light Pillars Over Latvia (2009 Jan 12)

[Don Stoner]

There were many good observations posted (and also many entertaining ones): First, logmark's observation that the pillars exist on a vertical plane between the source and the observer. Next, chadair's that the stars provide angular scale. And, of course , PiTHON's detailed observations. Finally, Stephen McDonald's that the "top" was closer than the base. Just how much closer might be a surprise! First the math and physics: light passing through horizontal rod shaped-hexagonal ice crystals refracts vertically at about 22 degrees (the Lowitz angle - from the geometry + Snell's law + 1.31 index of refraction for ice). When the distant sun is involved, this means the effect is observed 22 degrees from the source; but when the source is inside the crystals (like here) there is a whole family of paths which involve a single 22 degree downward refraction. This family describes an arch on logmark's vertical plane with one leg at the source and the other at the observer. The top (peak) of the arch is seen (by the observer) 11 degrees above the source (starts up at 11 degrees, then gets bent down 22 to -11). The arch is a circular arc which is comprised of 44 degrees (360-(180-22)*2, from the geometry) worth of a full circle. The tail-end of this arch hits the observer at the full 22 degree angle (so nothing an be seen above 22 degrees). Since chadair has placed the bottom of the fan at 17 degrees (well past the 11 degree middle) the "top" of the fan appears to be coming "up" on the full 22 degrees- putting it right next to the camera lens. (This effect would not be observable with stereoscopic photography because the "mirage" moves with the observer.) The fanning must be a ground-level effect! Some "bloom" can be seen around the sources suggesting dust or possibly moisture; but since light from all of these different optical paths must pass through this same ground layer, we need another explanation. I'll suggest that the crystals might be "fogging up" (would require exactly 0 deg C.) or that they might be losing their horizontal orientation near the ground. The low temperature probably falsifies the former. What does the "winner" get? I hope it's a cool APOD calendar! -Don Stoner (B.S. physics)

[Don Stoner]

Sorry, I need to learn to write: The source is inside the "cloud" of crystals - Not inside the crystals themselves. -Don

[Don Stoner]

One more thought: The crystals near the ground might be dirtier (4 mph wind with dust) - degrading their optical quality. The other paths (with the exception of the source "bloom") miss these low crystals. -Don

[Don Stoner]

When I plotted it (.gif should be attached) the fan at the top started looking more like simple airborne-dust scattering. The lower the path (both "top" and "bottom" of the pillar), the more near-surface dust it has to pass through and the more likely it will be scattered. -Don

[Don Stoner]

... but now it's starting to look like the "focused" section ought to be eleven degrees away from the source - where the highest and cleanest path is. (There goes my calendar.) Can I get a re-measure chadair? -Don

[logmark]

I'd like to point out that we are talking about evidence of light pollution. Lights that point downward or are otherwise shielded to direct their lighting only toward the ground do not produce light pillars. Only those light sources which permit light in strong sideways or upward directions give us these images. As astronomers (or buffs) ought we not be looking at ice pillars as evidence of offending light sources?

[neufer]

I'd like to point out that we are talking about evidence of light pollution. Lights that point downward or are otherwise shielded to direct their lighting only toward the ground do not produce light pillars. Only those light sources which permit light in strong sideways or upward directions give us these images. As astronomers (or buffs) ought we not be looking at ice pillars as evidence of offending light sources?

The Baltic States are far from being the worst offenders:

[mikedel]

Sometimes one tries to hard to see something or think to hard. I try to keep things simple. To me I believe that the fanning at rhe top is just weather related. It could be that as the ice crystals are falling they are in all different shapes and size. Size being the key word. The smaller crystals meltt before they can hit the ground, while the larger ones continue further on down. So the columns don't fan out at the top but rather the column narrows as the crystals fall. Fantastic pix though.

[kb9hpq]

I would suggest that as the base of the cloud becomes more dense the light is diffused over a broader area.

[mihondo]

This morning here in Midland, Michigan, we saw this effect!
Temperature = 0 deg F, Wind speed: 0 .

[nick.rassenti]

I'm am an architect.

I can't help noticing that the fanning pattern occurs primarily over what seem to be streetlights and not other light sources.

Street lights have an unusual light distribution pattern very different from simple point sources. They are designed so that in an array they will spread light evenly along the street. You will notice that when you walk directly under a street light that the light level actually falls off a bit, as if the luminaire has a blind spot. Maybe this has something to do with the fanned light pillars imaged by the ice fog.

[neufer]

I'm am an architect.

I can't help noticing that the fanning pattern occurs primarily over what seem to be streetlights and not other light sources.

Street lights have an unusual light distribution pattern very different from simple point sources. They are designed so that in an array they will spread light evenly along the street. You will notice that when you walk directly under a street light that the light level actually falls off a bit, as if the luminaire has a blind spot. Maybe this has something to do with the fanned light pillars imaged by the ice fog.

----------------------------------------------
. _The Adventure of the Six Napoleons_
........................................
"I wish to call your attention very particularly to the position
of this house, in the garden of which the bust was destroyed."
.
Lestrade looked about him.
.
"It was an empty house, and so he knew that
he would not be disturbed in the garden."
.
"Yes, but there is another empty house farther up the street
which he must have passed before he came to this one. Why did he
not break it there, since it is evident that every yard that he
carried it increased the risk of someone meeting him?"
.
"I give it up," said Lestrade.
.
Holmes pointed to the street lamp above our heads.
.
"He could see what he was doing here, and he could not there.
That was his reason."
.
"By Jove! that's true," said the detective. "Now that I come to
think of it, Dr. Barnicot's bust was broken not far from his red
lamp. Well, Mr. Holmes, what are we to do with that fact?"
.
"To remember it--to docket it.
We may come on something later which will bear upon it."
----------------------------------------------

[Trode]

This is one "cool" image . As one above has mentioned I to imediatly thought of Lightning "sprites" though there is little of that phenonama to apply here . I do agree with the above suggestion comparing the beam spread to "spotlight spread" pattern of stage lights . The light energy of the beem is concentrated ,hence brightest, where narrowly focused . Then the intensity drops off as one would expect with the spread/difusion past the focal point ....Trode

[Myriam]

hello!
I red only part of the readers proposals but my gess is more a camera optical effect than an atmospheric optical one.
When working on the picture with paint shop pro before joining you, I saw that the picture was shot outdoors through a stone window (no glass), see left side.
However, the camera was neither parallel to the first frame plan (right side is not so sharp than left one) and maybe nor to the pillars plan and this angle could have produced a sort of fish eye effet. The light are beginning to diverge where the "stone" frame is converging. the deepfield between the stone frame and the pillars and the difference in scale may explained the different angles.
What did the photograph really saw ?
Will we have a feedback about the more credible option ?

I take this oppotunity to thank APOD team for the always marvellous and challenging daily pictures.
Myriam . Belgium

[Aviator67]

Notice the convergence point toward the top of the image indicating considerable perspective distortion. This image was apparently shot through a window and I would agree with previous posts, perhaps a car or a train. The influence of a curved surface between the camera and the lights shouldn't be overlooked.

[james]

Unusual Light Pillars Over Latvia 2009-01-12

My best guess is that the light refracting ice crystals are dispersing the light above a focal point where the column is no longer visible from the camera's vantage point. A view from a higher elevation should confirm this guess.

The ice crystals are probably shaped much like a lens and focus the light straight up, but imperfections in the lens cause some of the light to be refracted horizontally. Multiply this effect by the number of ice crystals hanging in the air and you should see the phenomena shown.

The ice crystals are refracting the light horizontally, which is why the column is visible to the side. Much of the light is also being directed straight up through the transparent ice crystals to a focal point above which there is not enough light being refracted horizontally downward to be visible to the camera. The arcs above the focal point are where the ice crystals are refracting the horizontally refracted light from the column downward enough to be visible to the camera.

Again, this is my best guess.

Blessed are the Cracked for they let in the Light
james

[Myriam]

Hello, it is me again with a another approach than yesterday and closer to some readers' comments .

I thing that the light pillars' shape could be due to a spherical aberration of an imperfect lens situated inside the light pillar (top of the street lampin glass? ice cristals ) and pictured externally, meaning not as circles as it is currently the case when the camera/telescope lens situated inside the primary beam of light is to be blamed.
see
http://commons.wikimedia.org/wiki/File: ... tion_2.svg

when you rotate the bottom scheme of the wiki picture 90°left, it corresponds very well to the Latvian pillars, in particular the white one which shows the lighter part in the middle of the top of the beam of light and the read beam with its enlarged bottom part.

Myriam . Brussels

[Don Stoner]

I've cleaned up my earlier suggestion which was posted here, added some pictures and links, and generally tried to make it easier to follow. I've turned it into a web page which is now posted here. - Don Stoner

[pilotmba]

This is my first time for something like this.

I have read most of the other comments. I think several people are quite close to the correct explanation.

First, the two brightest pillars are not the only ones. The image has many verticals of varying degrees of brightness. As to the two most prominent, it's possible that the source lights have a strong "lobe" pointing upward. But the real question is what causes the opening-out of the beam towards the top. Again, I think several respondents were close, especially in pointing out that the beam seems to narrow to a focal point before it flares out. My refinement to the discussion is to point out that as the beams go higher, our visual perspective becomes more and more "underneath" the beam, so optically the flare appears to get broader. Light's straight-line travel continues until scattered by something -- air molecules, ice crystals, etc. So the scattering by the ice crystals is of course a probable cause.

GW

[ianrusselluk]

My guess is that it's a sub-zero cloud of flat, hexagonal ice-crystal platelets. Held in a stable condition below a layer of very slightly warmer air as often happens on a clear, frosty night (at least one star seems to be visible in the photo). In such still conditions the flat crystals would tend to orient themselves horizontally as they very slowly fall downwards through the air. This orientation would account for the vertical pillars of reflected light, as most light is reflected at 90 degrees to the orientation of the crystals.

It looks to me as if the 'trumpet-shaped' top of each pillar is near the upper boundary of the air-mass containing the ice crystals. This could be due to the horizontal orientation of the crystals increasing uniformly with increasing distance from the upper edge of the ice-cloud. The higher the proportion of non-horizontal crystals in the cloud, the wider the 'pillar' of reflected light.

One would expect crystals at the top of the cloud to be less oriented because these would be more recently-formed ones that are just beginning their slow gravitational descent. The stillness of the air would account for the mathematical perfection of the trumpet-shaped curves. If my guess is right, these curves are actually graphs showing the proportion of horizontally oriented crystals against altitude.

Ian Russell
http://www.interactives.co.uk

[alter-ego]

Hello,

Here is my two-bits towards understanding the light pillars - I've only read a couple of replies but I couldn't help diving into HaloSim to uncover halo behavior, especially related to light pillars

Long story short: Ice crystals have it. To first order, a combination of ice plates and horizontal columns can simulate the Latvian light pillars, even to the pronged flares. No lensing or optical distortion needs to be introduced. The angular dimensions of the simulation are also close to the real pillar.

The link below shows the HaloSim simulation and the light pillar image that has Cygnus stars & Vega visible. The angular scales for the picture and the simulation are matched (within ~10%). Based on the visible stars in the picture, the HaloSim scale was set via the zoom. FYI, the time of the exposure was ~11:45pm Latvian time (don't exactily know the horizon).

http://picasaweb.google.com/okubet/APOD ... 587538/img

Details & Discussion
Here's the breaout of the (thankfully) simple model: I used 20% Flat Plates & 80% Hexagonal Horizontal Columns. The plates were needed to get pillar extension, and the columns were needed to create the structured flare. The trick to the structure is to reduce the "c/a" aspect ratio in HaloSim. By letting c/a=> 1 (or so), the transition from 3-prong structure to a more uniform, filled in flare. Also, it looks like the HaloSim source size wants to be <0.5deg. I used an 0.4deg source diameter at -10deg elevation. There is quite a bit of "play" between the halo parameters, but the dominant features of the light pillars required two kinds of ice crystals to get close.

The simulation source is ~6 degrees below the actual source; the best I can tell. Given that I'm using HaloSim to model a totally different source than the sun (both in divergence and subtended source angle), I was totally surprised how close the simulation is to the picture. However, I was not surprised about the trend for source elevations to be <0 deg because the light reaching our eyes (from a light pillar) is reflected downward, therefore, light incident on the ice crystals is upward going (i.e. source is at negative elevation).

In conclusion, methodically using HaloSim to search for light-pillar behavior, I landed on a relatively simple solution that is very close to the real Latvian light pillar. Close enough, in fact, that I believe ONLY ice crystals and their orientation are needed to explain the Latvian examples. Nothing more complicated is needed for the primary ice pillar features. However, close inspection of the ice-pillar pictures to reveal fine structure detail not simulated. Also, variation is ice-crystal densities between the observer and the source are not modeled in HaloSim. Density differences could affect the size and/or shape of the pillars, and as I said, I don't expect the simulated source location to be correct.

Is the model a lucky coincidence? I don't think so. It's those phenomenal ice crystals!

[paulogram]

I believe it has to do with the density of the ice crystals ... I would think that the ice crystals slow down in the thicker air closer to the ground and therefor bunch up (akin to a layer of fog blanketing the ground), being more dense and reflecting a thinner column of light down low... but at a certain height above the lower ground layer of dense air & crystals is a layer where the ice crystals become less dense in a gradient fashion ...perhaps temperature gradient assisted by a breeeze at the higher altitude (closer to the ground the air is more still) ... the less dense the ice crystals ... the wider the reflection pattern causing the fan shaped top of column

[pilotmba]

I'm afraid I have to disagree with "paulogram." The denser the dispersion medium, the greater the scattering -- not the other way around. I still stick with my theory that the density of ice crystals may be quite uniform throughout the vertical cross-section, and that the fanning of the light beams towards the top is a matter of our perspective changing as we look towards the top of the columns. Also, I am not sure that the geometry of the individual ice crystals is a variable, as long as the geometry is fairly similar bottom-to-top. I can reproduce the same effect on a misty night with a well-focused flashlight. Try it. Take a flashlight out on a foggy night and shine it in any direction. The beam seems to open up non-linearly the farther it gets from the flashlight. Dispersion media will do that, since it has more and more of an effect the farther the beam penetrates through it.

GW

[tidman]

give that the light sources are unchanging, and ice particals are variable. seems the "reflections" are different due to a special arrangement of the reflectors (ice crystals). might this be a demonstration of an electro-magnetic field present, where in the charges on the ice (water is a polar molecule) formed in a precise temp, and pressure, causes them to align in that layer in a uniform pattern...resulting in a uniform reflection to the viewer. The "flare at the tops would be the level where the field is weakening or the crystals are not formed in their polar arrangement.
Tidman

[dennis]

I think the fanning is due to a moisture gradient falling off rapidly moving upward from where the fanning starts.