APOD: A Triple Sunrise Over Gdansk Bay (2009 Aug 04)

[polymath]

Vertical-convection-current collimated and axially oriented condensation products in the stratocumulus clouds act collectively in high magnitude exponential numbers as prisms. The Sun's light is radially polarized by the millions of kilometers of intervening distance in the vacuum of space, effectively arriving in a linear beam.

I know the difference between ad hominem and argumentum tu quoque.

[RaginCajun]

I just did a quick search of Flickr and found these two pics... In both cases the photographers clearly state that the image results from a window effect. I haven't looked at all the images from "double sunset" and "double sunrise" searches. There may be more examples.

http://www.flickr.com/photos/kiewee/422958365/
http://www.flickr.com/photos/siriustar/471806578/


Apologies if these have already been posted. I've been checking in on the discussion but haven't read every post

[apodman]

Vertical-convection-current collimated and axially oriented condensation products in the stratocumulus clouds act collectively in high magnitude exponential numbers as a prism. The Sun's light is radially polarized by the millions of kilometers of intervening distance in the vacuum of space, effectively arriving in a linear beam.

I know the difference between ad hominem and argumentum tu quoque.

That's really good stuff. Very entertaining.

[GliderPilot]

The angle between a pair of internally reflected images is 1/3 the angle between the reflecting plates.

My calculation was that the apparent angle between the images is twice the angle between the plates, not 1/3, but regardless, there have been enough photos posted through double pane windows to show that the separation of the sun images is entirely within the realm of normal separation due to variations in the window. That leaves the only good argument for the atmospheric phenomenon that the observer was interested enough to have noticed this variation over the surface. If it was only a casual "Gee - what's that, let's take a photo" it's quite possible that a variation over the window was mistakenly seen as a variation over time.

The argument against atmospheric phenomenon remains strong - horizontal density variation sufficient to cause this magnitude of image separation would be quite rare - if it ever occurs.
My own analysis is converging on the window pane explanation.

[Chris Peterson]

Vertical-convection-current collimated and axially oriented condensation products in the stratocumulus clouds act collectively in high magnitude exponential numbers as prisms.

References? Sounds like so much gibberish to me. FWIW, the highest local horizontal pressure variations observed on Earth occur at hurricane eyewalls, and can reach perhaps 15 kPa in category 5 storms. That's the equivalent pressure shift from a 1000 m change in altitude. In the APOD, we are seeing the Sun through a gradient index medium caused by the entire height of the atmosphere (with little vertical distortion), and there is no sign of anything like an intense tropical storm to create a horizontal density change. And assuming a 15 kPa horizontal stepwise shift in air pressure, you might want to calculate the two indices of refraction and see if a 0.75° shift is possible. Or you might want to figure out just what change in index would be required for that amount of deviation, and what pressure differential would be required. While you're at it, look at the dispersion of our atmosphere, and explain why this stepwise density gradient isn't resulting in images that show some spectral separation between the left and right sides. (The calculations aren't very difficult... any polymath should be able to trivially perform them.)

FWIW, typical local pressure variations from ordinary winds and updrafts are around 0.25 kPa.

The Sun's light is radially polarized by the millions of kilometers of intervening distance in the vacuum of space, effectively arriving in a linear beam.

The Sun's light is demonstrably not radially polarized. Traveling through vacuum doesn't result in any EM becoming polarized (not linearly, circularly, or radially). There is no such thing as a "linear beam". In short, this entire sentence is a mixture of nonsense and factual error.

[polymath]

Res ipse loquitur. Stratocumulus clouds over land, open expanse of sky above water. Sunrise, water warmer than the land, convection currents carry warm moist air aloft and it descends over the land condensing into clouds. The foreground stratocumulus clouds' edges' crispness indicates light horizontal winds at cloud height. Gentle downdrafting currents orient a portion of condensation products vertically, light horizontal winds orient a portion of condensation products horizontally. A sufficient quantity of condensation products are oriented ideally to project multiple refractions. A delicate balance, I'm sure, probably why multiple mirage sunrises are so rarely observed.

I believe atmospheric density gradation is a red herring because the available evidence indicates to me the refraction is caused by condensation products.

Sunlight's radial polarization is an obsever consequence of relative location. Sunlight radiates out omindirectionally. The portion that arrives at an Earth observer's location is a nearly linear beam in most respects. Although if it had continued to infinity, the individual rays would be separated an infinite distance.

[Chris Peterson]

Res ipse loquitur. Stratocumulus clouds over land, open expanse of sky above water. Sunrise, water warmer than the land, convection currents carry warm moist air aloft and it descends over the land condensing into clouds. The foreground stratocumulus clouds' edges' crispness indicates light horizontal winds at cloud height. Gentle downdrafting currents orient a portion of condensation products vertically, light horizontal winds orient a portion of condensation products horizontally. A sufficient quantity of condensation products are oriented ideally to project multiple refractions. A delicate balance, I'm sure, probably why multiple mirage sunrises are so rarely observed.

I believe atmospheric density gradation is a red herring because the available evidence indicates to me the refraction is caused by condensation products.

It appears to me you don't even understand the difference between refraction and reflection. If you think this is caused by water droplets, describe the internal reflection geometry that could produce a secondary image 0.75° from the primary. And explain why dispersion of the light passing through the droplets doesn't produce some spectral shift.

A real world example of an effect caused by internal reflection in water droplets is a rainbow- which does not produce an image of the Sun, and does show spectral dispersion. A real world example of an effect caused by reflection from ice crystals is a sundog, which does not produce an image and does show spectral dispersion. There are no identified halo types that operate over such a narrow angle, other than simple scatter mechanisms- which do not produce images.

So far, nobody has remotely made a case for a plausible atmospheric effect that could cause two undistorted ghost images of a rising Sun.

Sunlight's radial polarization is an obsever consequence of relative location. Sunlight radiates out omindirectionally. The portion that arrives at an Earth observer's location is a nearly linear beam in most respects. Although if it had continued to infinity, the individual rays would be separated an infinite distance.

You are completely incorrect here. Sunlight is not radially polarized. Your definition of a "linear beam" is a truism. Physically, there is no such thing as a light ray; these are simply tools for the mathematical treatment of optics.

[polymath]

I've stated the case for what I see as a likely cause of the triple sunrises. My position has been attacked by browbeating point-by-point refutation arguments based on attacking my perceived stupidity. I've said my peace and rest.

[DinK]

EDIT: Don't bother replying to this post. I got a chance to test this out for myself, saw what the deal was, and posted another reply below. I'll leave the original post intact below though.

It's been fun reading the discussion here, but I think when people want to talk authoritatively about the direction and angle of reflections, we should all back it up with pictures to show people what we're talking about. It would go a long way towards convincing others we're right, or at least clearly showing them when they're wrong.

I'm particularly thinking of this argument on whether the reflections would be to the right vs to the left depending on the angle of the double-glazed door or window. Some people are claiming that the overall angle relative to the observer and sun makes no difference and that only the angle between the individual panes matters as to whether the reflections are to the left or right. It doesn't seem to make any sense to me, so I'd like to see a convincing diagram if it is in fact true.

On page 8 I see some diagrams from Naos supposedly proving this. They certainly show that the degree of separation depends on the angle between the panes, but not that the entire direction does. There are two images with the overall window angle the exact same. Obviously in one the lines are diverging while they're converging in the other, but their departure points from the window are still in the same direction relative to the original light beam. It seems to prove the opposite point. Thankfully on page 10 Naos posted some simulations with the overall window assembly in different orientations, which clearly shows how the orientation does affect angle. Plus of course there were several posts containing real-world images showing real reflections on either side of the light source depending on orientation.

For my last point I wanted to pick on geckzilla a bit. You sir posted the picard facepalm for people arguing against you, but on page 7 you thankfully posted an image illustrating the supposed behavior of these window reflections. Granted you hand drew it and the angles aren't exact, but seriously the ballpark directions are so off I don't know how you looked at it as an illustration of light bouncing off a window and said "yep, ready to post!"

I think the angle should be the same in the opposite direction, not half in the same direction...

reflections.GIF (8.81 KiB) Viewed 3064 times

[eachun]

In Sept 1997 I was anchored on the west side ov the island of Contoy just north of Cancun Mexico. saw a sunset where there were two suns seting at the same time untill they were about on third of the way down then only one.I don't know what caused it, we saw this with the naked eye only.Any ideas?

[bystander]

@DinK: See Dave Shaffer, previous page.
http://asterisk.apod.com/vie ... 31#p109131

[Chris Peterson]

I'm particularly thinking of this argument on whether the reflections would be to the right vs to the left depending on the angle of the double-glazed door or window. Some people are claiming that the overall angle relative to the observer and sun makes no difference and that only the angle between the individual panes matters as to whether the reflections are to the left or right. It doesn't seem to make any sense to me, so I'd like to see a convincing diagram if it is in fact true.

The raytrace diagrams show this to be the case. Some of the diagrams, showing a single ray coming in and bouncing around, are confusing, however.

There is absolutely no doubt that the angle you look through the glass has no effect on the direction of image separation. That is determined only be the angle between the glass panes themselves. If you don't like the diagrams that have been posted, what about the actual photographs demonstrating it experimentally? Or, you can trivially do the experiment yourself.

For my last point I wanted to pick on geckzilla a bit. :wink: You sir posted the picard facepalm for people arguing against you, but on page 7 you thankfully posted an image illustrating the supposed behavior of these window reflections. Granted you hand drew it and the angles aren't exact, but seriously the ballpark directions are so off I don't know how you looked at it as an illustration of light bouncing off a window and said "yep, ready to post!"

The conclusion is correct, but the image as drawn is not.

[Eamon Shute]

This shows how a light ray is affected by multiple reflection between two panes which are not quite parallel. Ray A goes straight through, ray B is reflected twice and appears to emanate from the left of the sun, ray C is reflected four times and will seem to come from even further to the left.

[geckzilla]

Dink, I was surprised by what I saw, too. These pictures only roughly illustrate what I observed with my experiment. You can pick my crude little beam of light (it's actually just a black arrow) apart all you want, but what I saw were reflections which would always "bounce" toward the direction that the glass panes converged. I don't offer any explanation why because I don't know why. I wish I had a better picture and a better explanation but I don't. I've tried figuring out the path that the light bounces but have failed.

In fact, when I originally made my drawing, I made it wrong. I drew the arrow going the way I logically thought it would go and had given them angles that looked a lot more realistic, like ball bounces. After I picked the glass panes up to make sure I was doing it right, I was surprised to see I was completely wrong. The only thing I can conclude is that what is actually happening isn't what I think is happening but I just can't think of why or how it really is happening.

[Chris Peterson]

In Sept 1997 I was anchored on the west side ov the island of Contoy just north of Cancun Mexico. saw a sunset where there were two suns seting at the same time untill they were about on third of the way down then only one.I don't know what caused it, we saw this with the naked eye only.Any ideas?

Such effects are fairly common. The usual cause is a reflection from clouds. This isn't a specular reflection, so you don't actually see an image of a second Sun, but to the naked eye you see a very bright spot that can appear quite similar to a second Sun. A sundog can sometimes resemble a second Sun, and sometimes a cloud will bisect the Sun making it appear as two. There are lots of pictures on the Internet showing double sunset illusions.

[luchio]

Here are photos i made 5 minutes ago to show how easy is to get 4 moons using a gigital camera and a window with two glasses:

4 moons (extra one from camera lens)

4moons.jpg (26.08 KiB) Viewed 3009 times

Difrent angle beatween camera and window:

dirfent_angle.jpg (27.32 KiB) Viewed 3005 times

To get all reflections and moon in one horizontal line i need to place camera higher, but my tripod is too short so picture is blured:

blured.jpg (23.95 KiB) Viewed 3000 times

Maybe this could help imagine that reflections of the suns can be anywhere on the pictures...
I didn't rotate that pictures. As a proof there is Jupiter on the bottom right.

PS. Of course, all the reflection were visible to the unaided eye (through windows glass)

Luchio, Poland

[kamoses]

Ok, my explination of why orientation doesnt matter (look out everyone...this might get rough ):

First of all, this is only true of objects at infinity. The below diagram shows an object at a FINITE distance situated at where the colored lines convulge. The object has light going out in all directions, but we're just going to look at directions red, blue, and green. At the bottom is a circle, which represents your eye.

Path red: This is true of both a finite and infinite object. the light enters at an angle, reflects to the outside pane, and back again. Based on geometrical theorems (cant remember which) lines will be parallel if windows are parallel because angles must add to 180 deg. The thing to note here is only the original line makes it to the eye. All other lines miss the eye to the right, so we don't see the object translate right via the blue line.

Path blue: This line leaves aimed to the left side of the eye, and the primary light misses the eye. The light reflects in the window to the RIGHT side, and then reflects back to the eye. the light will continue to reflect down the window, but all other lines (not drawn) will miss the eye as with the red line. Because some light was lost each time the light reflected off a window, the image will be slightly dimmer than the primary red image.

Path green: This line comes out at a steeper angle. It acts the same way as line blue, but has to bounce back and forth many more times than line blue before it reaches the eye, so its dimmer. (I didnt draw all green reflections to make the picture less cluttered)

So for a finite object, we see reflections oriented dependent on the orientation of the window.

For INFINITE objects, the green and blue lines do not exist. Unless of course, the sun were to take up half the sky, course then we'd have other problems. For our sun, all lines are oriented in the direction of or extremely closely to the red lines, so all we see is one image.

Make sence?

[DinK]

Chris and everybody else,

You guys are right. The overall angle doesn't matter and it is indeed just a matter of how parallel the window panes are. Now that I'm at home, I have a sliding glass door and a flashlight with which I can easily conduct little experiments. The direction of the ghost images clearly depends on which part of the door I view the flashlight through and not the angle of the door relative to me and the light. So science and/or data wins again.

I think the vector diagrams can be misleading at a glance since it's easy to assume the viewer sees the image where the vectors originate, rather than where they converge, or something like that.

[DinK]

To address the original point of this thread: with any doubt about the direction of the ghost suns removed from my mind, I firmly think this is just a reflection in the glass. I think the visual indications of this are numerous and have all been discussed already. Furthermore if the photographer was stationary, whether sitting or in bed in this room, I think it would have been easy to see the suns and reach for the camera without noticing the reflections change much. If the person taking the picture was moving around the room, they probably would have noticed it was just reflections in the glass. For that matter they probably would have gone outside if they were already walking around.

[astrobug]

OK, After having spent way too much time reading these posts, I am firmly convinced that this is a reflection in window near the photographer, rather than some new atmospheric phenomenon in the distance. Rather than rehashing all of the prior arguments for and against the window hypothesis, I will add a separate(and to me, convincing) line of reasoning. In any sun mirage or related I have ever seen or heard about (true mirage, green flash, sundogs, etc), the mirages/reflections are always about the same _brightness_ as the principal (ok, sundogs are a little dimmer, but smeared out over a large area). In the images in question, the vase foot on the 'original' sun is about as bright as the sun proper (illustrating a well-known atmospheric effect), but the second and third 'suns' are markedly and progressively dimmer (the second 'sun' sporting its own vase foot of similar brightness). Window glass has a reflectance of about 4% per surface, so assuming a double-pane window is responsible, each successive reflection should be ~ 1000 times dimmer than the previous one (.04*.04 = 0.0016). This agrees pretty well with casual inspection of the photo. You can't trust the accuracy of the measured pixel values for photometric purposes because it's a .jpg file and the camera's response to light is likely non-linear to start with.
So, to further constrain what an atmospheric phenomenon would have to do to create this image, there would have to be two separate, natural 'mirrors' out there (one for each image), one dimming the image by a factor of about a thousand, the other by a factor of about million. So to summarize, any such natural 'mirrors' over the horizon would need to be very large (so that they would cover at least 1/2 a degree, which would be 8.7 meters of diameter for every kilometer of distance!), optically flat so as not not induce distortion/changes in size according to magnifying power, and with just the right reflectance/transimittance values. The first 'mirror' would need to have a reflectance of ~0.1% (transmittance of 99.9%) and the second's would need to be ~0.0001% (transmittance of ~99.9999%, which is better than any anti-reflective optical coating I've ever heard about...I'd love to have such a coating on any/all of my lenses!). So it's either a window reflection or the aliens have landed!
-Brett

[bartderluyn]

i think, if there was reflection far away,
the double and triple reflection would be not so te same,
so the source of the reflection must have been close to te lense or the eye,
but I see no evidence of a window

[dockwatcher]

So the building sets at 115 degrees (and the window if closed) and the sun was at maybe 75. So wouldnt the reflections be on the right side?

[Chris Peterson]

So the building sets at 115 degrees (and the window if closed) and the sun was at maybe 75. So wouldnt the reflections be on the right side?

Review the past posts. The angle of the glass pane with respect to the source and viewer has no impact on the side that ghost reflections are seen.

[dockwatcher]

How about if I just look at some of the previous posts.

If light comes in from the left it bounces right. if you move the camera the image moves.

The moon shots show a clearer secondary image then the primary. Ours show considerable distortion (mostly horizontal and on both sides).

I don't know why it seams so obsurd that somebody could have droped and bent their 11 element in 9 section (one active) lens and discoverd it takes cool opictures.

[apodman]

I don't know why it [seems] so [absurd] that somebody could have [dropped] and bent their 11 element in 9 section (one active) lens and [discovered] it takes cool [pictures].

If that were the case, wouldn't everything in the picture (not just the sun) be affected?

Before someone else asks, doesn't the same question apply to the window glass? We have already read the answer about the attenuation of the sunlight with each subsequent reflection off the window panes. The images of other objects are also reflected, but (unlike the sun) they don't start off bright enough for their reflections to be visible after attenuation. The camera lens system is designed to minimize internal reflections in the lens train almost to zero. I suppose you could misalign and displace the lens elements in some manner to produce the same reflection and attenuation as window glass, but I ask you which is more likely: a random alignment resulting from a dropped camera acting like window glass, or window glass acting like window glass?

And, taking the photographer's word which we have shown no reason to doubt, remember that the same reflections were visible to the naked eye before the photographer used the camera to record them.

One more related question ... Which is more likely: an unknown (and so far unsupported by a convincing scientific explanation - or by other documented examples of the same effect) atmospheric phenomenon acting like window glass, or window glass acting like window glass?