Ghost Moon?

By Sean Carroll | October 25, 2010 9:50 am

Does anyone know about this phenomenon? My friend Benson Farb, under the charming misimpression that I am some sort of astronomer, sent me the following image, taken by his uncle Henry Farkas, MD.

Ghost Moon

That’s the Moon on the right, somewhat overexposed. On the left is another image of the Moon — substantially dimmer. So what is going on?

Consulting the Google, I was able to find multiple examples of similar phenomena, but no explicit explanation: see here, here, here, here, here, here. My first guess was that we were glimpsing a giant Death Star that had been hiding behind the Moon, but upon further thought I regretfully concluded that it’s unlikely we would have alien invaders clever enough to build a Death Star but sloppy enough to reveal it prematurely like that.

Actually there is only one sensible explanation: some sort of lensing/reflection phenomenon that is giving rise to multiple images. The two obvious culprits would be the camera lens itself, or the atmosphere. But Henry took the picture in the first place because he saw the ghost image with his naked eyes, so the camera lens is out. Atmosphere it is! This is somewhat corroborated by the fact that different exposures show different separations between the images — something that could be explained by changing atmospheric conditions.

Ghost Moon 2
Ghost Moon 3

The atmosphere, whose layers can have very different humidity and temperature, can be a very effective reflector and refractor. Here is an image of a “sun pillar,” to show how dramatic the effects can be.


So I’m pretty convinced that the atmosphere is to blame. On the other hand, it’s a little funny that the images aren’t vertically aligned, which is what I would naively expect. And this wouldn’t be the first time that my lack of real-world knowledge steered me dramatically wrong. Anyone familiar with this phenomenon?

  • ian

    That’s no moon.

  • Dan Fischer

    This has all the marks of an internal reflection of the lens system to me. Easy test (esp. since several pictures exist): Is the “ghost” always in a spot which is a point reflection of the Moon’s position relative to the center of the field of view? If so: case closed! Bonus: the top picture here shows such an internal reflection of a partially eclipsed Sun, which also highlights the geometry involved.

    • Sean

      It can’t be an internal lens reflection if it’s seen with the naked eye.

  • Steven

    What if the naked eye is looking at the moon through corrective lenses? Those could also introduce a similar reflection. I’ve noticed my glasses producing extra images of high contrast sources like the full moon against a dark sky.

  • Pingback: Tweets that mention Ghost Moon? | Cosmic Variance | Discover Magazine --

  • ENT-TT

    It appears as though Luna has been tidally locked by a red dwarf.

  • spyder

    I suppose you could investigate some of the variables. Are the images of the ghost moon taken as the moon is coming up and/or going down versus directly overhead? Does the camera image accurately represent what the viewer sees with the naked eye, not wearing glasses?

  • amphiox

    “My first guess was that we were glimpsing a giant Death Star that had been hiding behind the Moon, but upon further thought I regretfully concluded that itโ€™s unlikely we would have alien invaders clever enough to build a Death Star but sloppy enough to reveal it prematurely like that.”

    Ah, but the Tarkin Doctrine emphasizes terror over utility. An early reveal to a victim with no hope in hell of stopping the thing would not be inconsistent. . . .

  • Jonathan Lubin

    It would have helped to know the date, time, and location of the shot, as well as the weather conditions at the time, and the model of the camera. The three .jpg images aren’t even of the same size, in pixels (download the images and check the files!). Eyewitness account of the same phenomenon being seen without the camera notwithstanding, I lean heavily to the in-camera explanation.

  • alan beever

    i saw that moon at 07:00 yesterday morning .it was a very clear morning with full moon .I said to the wife we got two moons very weird.

  • Stephen P

    I think this could be caused by a thin, sharply-defined layer of mist/cloud (or just possibly dust?). The second moon is the illumination of this layer.

    In principle the image would appear directly between the observer and the moon. However the most likely cause of such a layer would be the interaction of two layers of air: one cold and the other warm and moist. In this case the different densities of the two layers could cause the image to be refracted away from its nominal position.

    What I don’t understand though is why the ghost moon is smaller than the moon.

  • Bob

    You sure it wasn’t taken through a glass window. Multiple reflections in the glass will do it. Normally you can’t see them unless there is a lot of contrast like the full Moon in a dark sky.

  • Henry Farkas, MD

    I’m the guy who took the photos. I was outside, not looking through a window, and I wasn’t reading, so I wasn’t wearing my reading glasses. When I saw the image with my eyes, the ghost image was partly overlapping the actual moon image and was at an angle to it. The images were taken only a second or so apart because I have my camera set to take three images every time I press the shutter release, or whatever they call it these days with digital cameras. The first image is taken at whatever exposure the camera’s computer thinks is correct. The second image is 2/3 of an f stop darker, and the third is 2/3 of an f/stop lighter than the camera’s computer thinks is correct. The camera was oriented horizontally just the way the photos are displayed above. The top image is cropped a bit and darkened a bit to get rid of the surrounding glow, but the other two images are not edited at all.

  • Henry Farkas, MD

    On October 25th, 2010 at 12:03 pm, Jonathan Lubin asked for some information about the details of the camera. Here they are:

    The camera is a Canon PowerShot SX 10 IS superzoom.
    The focal length was zoomed to 100mm. This is the 35mm equivalent of a zoom to 589mm. It was hand held so clearly the image stabilization kicked in.
    The photos were taken on Oct 21st, 2010. The weather was severe clear.

    The first one registered at 9:07:03pm. The shutter speed was 1/20th of a second and the aperture was f/5.7. This first shot was the one where the ghost image is touching the real image.

    The second image registered at 9:07:04pm and the third, the one I cropped, registered at 9:07:05pm. Although I thought I had the camera set to record a regular image, a darker image, and a lighter image, all three images, now that I look at the metadata, had the same shutter speed and aperture as the first image.

    I have to reiterate, there was no glass of any kind between me and the moon when I saw the ghost image. There is one detail that I should mention. Cameras don’t see as well as eyes do. To my eyes, the ghost image did look like it was exactly the same size as the actual moon. It was just not anywhere near as bright as the actual moon. You have to remember that my camera isn’t all that high quality. It’s a point and shoot with a zoom lens.


  • ChH

    Looks like what was left of The Great Evil after Milla Jovovich got done with it.

    But Seriously – Dr Farkas – what was your latitude when this was taken?
    Also, have you ever had radial keratotomy or other surgery to correct your vision?

  • Dan Fischer

    @Henry: Did you do the test I proposed in comment #2 already? Must be done on uncropped versions, of course, i.e. you have to know the center of the FOV precisely. Given a sufficiently bright light source (the Moon would still do it tonite) and identical or similar camera settings, the effect should also reappear. The non-Moon-ly color of the reflection is caused by lens coatings, by the way; this topic comes up on astronomy web sites quite frequently.

    Now regarding your visual impression: How long it did last and was that “ghost” changing its distance and/or angle from the Moon while you saw it? My hypothesis predicts that it didn’t, as in my view what you saw was a different (physiological) phenomenon in your eye(s) – where a lot of optical ‘elements’ are present that can scatter light into glare effects in an asymmetric fashion – which would follow different geometrical rules than the internal lens reflection.

    Just some thoughts from a seasoned amateur astronomer who has seen lots of visual as well as photographic artefacts over 30 years …

  • Jason A.

    I’ve seen identical images many times in pictures I’ve taken that were never visible with the eyeballs. The separation from the moon can be changed by positioning of the moon in the field of view (always symmetric about the middle). Different degrees of brightness and colors is possible (I was once fooled into thinking I discovered a comet by a green reflection from a bright star), and changing lenses can change that. I always photoshop it out. So basically, it can be caused by the camera optics. Claiming it’s visible to the eye is just an anecdote ๐Ÿ˜›

    Check the pictures and see if the reflection is symmetric with the moon. If so, what’s in the pictures is an internal reflection. Which doesn’t mean that he didn’t see anything with his eyeballs, but whatever he saw would not be the same thing as what’s in the pictures.

  • Euclid

    Check out
    for the Flat Earth Society take on this.

    Perhaps it’s the Anti-Moon/Shadow object theorized to cause eclipses in Flat Earth theory?

  • Henry Farkas, MD

    I was in the parking lot outside my brother’s condo. According to, I was at
    Latitude, Longitude
    39.363969,-76.468644 Just east of Baltimore city.

    I can’t duplicate the photo with my camera tonight. It’s cloudy outside. But I have to tell you, I take photos of the moon every so often just to see how steady I can hold my camera (ah, the glories of digital photography, where you can take pictures for free) and I’ve never seen this happen with my camera before. I’m not sure what test I’m supposed to do about that FOV (field of view?) thing that was mentioned in response #2.

    Jason A says that if it’s an internal reflection, the ghost image and the moon will be symmetric with the center of the image. You can’t tell with the cropped image, but the other two images are not edited at all, and the ghost image is not symmetrical with the actual image around the center of the photo.

    When I was seeing the double image with my eyes, the ghost image wasn’t anywhere near separated from the real image. There was around a fifty percent overlap. And it wasn’t moving. I originally thought my eyes weren’t working together. That’s why I took the photo. Seeing double is definitely not recommended by the eye doctors or the neurologists. Vertical or diagonal diplopia (double vision) is a very bad sign.

    When I take photos with my camera, just like with a single lens reflex, you can’t see the image you’re photographing while the image is being captured. It’s not as if my camera has a mirror that flips up like in the SLRs. The camera company, Canon, could just as easily allow me to continue seeing the image as it’s being captured. They just don’t. This is a bit problematic when you’re capturing three images in a row since it’s difficult to keep the camera precisely aimed when you can’t see what you’re aiming at for two whole seconds. That’s why the image of the actual moon isn’t in the same spot in the two uncropped photos.


  • Andrew Cooper

    The internet source for this stuff is a website called Atmospheric Optics (, check through the site for all manner of strange things our atmosphere can do. If you can’t find the answer, send in the pic for a diagnosis, Les loves to collect pics of different phenomena.

  • Carl Brannen

    The only thing I can think of that can explain this is that it’s the moon illuminating a small circular / spherical cloud perhaps of reddish particles. This explains the smaller size, the red coloring, and the fact that it is moving.

    Ice crystal phenomena typically have a preferred angle. For example, we’ve all seen sun dogs and moon halos.

    By the way, my understanding is that a sun pillar is due to ice crystals aligned by gravity causing reflection / refraction rather than by refraction due to air density alone. I’ve only seen them in the presence of high (i.e. icy cold) clouds.

  • Dan Fischer

    @Sean (comment 3): In none of his comments Henry has confirmed that the thing he saw behaved in any way as the thing on the photographs, namely changing its relative position by several lunar radii within seconds. This is a giveaway that what he saw (eye glare from the bright Moon as the “weather was severe clear”) and photographed (lens flare, frequent in point & shoot cameras) is unrelated. Henry also stresses (comment 14) that the visual ghost moon had the same diameter as the real Moon and was – apparently constantly? – offset by 1/2 diameter. Which is not the case in any of the 3 photographs by a large margin. Finally regarding the idea (Sean’s original post) that atmospherics optics might be involved: I’ve attended a number of conferences on this topic, covering both frequent and rare phenomena – there simply isn’t one that involves a 2nd image of a bright object jumping around in its vicinity, let alone when the sky is clear.

  • Carl Brannen

    Ah, here we go. A little google-fu shows that this is a problem for amateur astronomers. Basic light pollution problem with clouds:

    “Faint stars disappear, only the Moon and the brightest planets stars remain visible. If the night is cloudy, the light pollution effects are striking. The clouds are red and bright and one could read the newspaper using the light reflected by the clouds. So much light is uselessly pumped toward the skies that the reflected light is enough for a decent illumination. I’m fairly sure that was not the intention here… The waste is enormous. At a fraction of the cost, we could have good design lamps with lower power which can provide the same illumination.”

  • Dan Fischer

    @Carl: Look at the exposure times and f numbers Henry gave in comment #14 – the (as I explained non-)phenomenon is orders of magnitude brighter than any light pollution-lit cloud could ever be (and there were no clouds anyway, as he said). Apparently “google-fu” doesn’t replace actual sky observing experience …

  • Brad H

    The last two images shown above were taken only a second apart?
    Surely it can’t be an atmospheric phenomenon then – it would require
    high altitude clouds to be moving in pretty spectacular ways to shift
    the image like that. And coherently too!

    It’s gotta be a near field effect. If not the camera optics, then some sort
    of ground layer fog, although I believe the weather was claimed to be clear. Could the temperature have been hovering around the dew point?

  • Dan Fischer

    As this is a physics blog I did … a physics experiment and just shot the waning gibbous Moon a few minutes ago, in very clear skies with a (different) point & shoot camera but also heavily overexposed. My own ‘ghost’ is there as expected, of a different shape and color due to the different zoom optics (and at a different position relative to the Moon in each frame, depending on where the latter was in the field of view). And yes, at the same time I saw a distinct ghost of the Moon with my own naked eyes, too, of the same size as the Moon, much fainter and overlapping. On the 21st, around full moon, such glare effects would have been even more pronounced. Case closed, I’d say.

  • Aaron Sheldon

    Is it related to Sun Dogs and high atmospheric ice crystals?

  • Navneeth

    27. Aaron Sheldon, this particular phenomenon is most likely not. But sun dogs are due to the presence of atmospheric ice crystals. Check out the link provided by by Andrew Cooper (#20).

  • chris

    it sure is gravitational lensing :-)

  • Reginald Selkirk

    I used to get ghost lights when I would put a UV Haze filter on my camera+lens. They would be inverted about the central point from the original/real light.

  • jd

    Atmospheric birefringence.

    Of course, camera filters/mirrors can produce the same effect if you wanted to hoax it.
    The moon will shift around depending at what angle the front element is to the subject.
    Remove front filter (s) or tilt them to correct.

    Whats telling to me is the colour of the second moon…

    I have seen this effect with the sun (sorry no camera).

  • michaelspierce
  • ChH

    It’s definitely not a sun (or moon) dog. They are always 22 degrees away from the light source due to the prism effect of ice crystals. Dr. Farkas’s pictures show a second image only to one side of the moon, varying between 0.3 and 2 degrees away, center-to-center.

    I agree with Dan Fischer – we are looking at a false 2nd image due to the camera optics, and Dr. Farkas also saw a false 2nd image with the naked eye due to the “optics” of his eyes.

  • Henry Farkas, MD

    Reginald Selkirk’s comment that a UV/haze filter could have caused the problem might be the answer. My camera typically wears its UV/haze filter all the time, and it was wearing it that evening. My theory is that a scratched UV/haze filter is cheaper to replace than the whole camera would be if the lens got scratched. These superzoom cameras don’t come with removable lenses like the digital SLRs do.


  • Jonathan

    The effect with the naked eye is likely a distortion effect caused by moisture in the eyes. This can often cause what appears to be a slightly separated double moon (not normally as much as in the photos). The different positions of the ghost image are probably because the camera moved slightly during the shot. Were these shots handheld or on a tripod?

    Sun pillars are caused by plate crystals lying roughly horizontally in the sky. The larger the deviation from the horizontal, the taller the column. Changes in density and moisture levels in the sky don’t change this. Refraction effects can cause distortion of the moon by splitting the colours (green at the top, red at the bottom) like in this image (, but this always happens very close to the horizon and is only ever a very small effect.

  • Carl Brannen

    “the phenomenon is orders of magnitude brighter than any light pollution-lit cloud could ever be”

    The photo was taken just NE of Baltimore, Maryland. It’s one of the most light polluted spots in the country. It’s Bortle class is 8 or 9. See for the light pollution map of Maryland and compare with the latitude / longitude given.

    In such areas, the light from clouds is sufficient to read a newspaper at night. And that light can be reddish orange. For example, see: “Such lights reflect off low clouds and color the sky reddish orange, says astrophotographer Jerry Lodriguss.”

    “(and there were no clouds anyway, as he said)”

    When you see something in the sky you need to take into account the outside chance that it is a cloud.

  • Carl Brannen

    You can see the cloud for yourself. October 21, 2010 9:07PM EDST is October 22, 2010 01:07 UTC. If you go to you can see for yourself the satellite images of the clouds over the eastern US at this time by setting the satellite to GOES East. (Try to do this before Thanskgiving 2010, as they only keep data for a limited time.)

    Set the times to run from say October 21, 11:00 UTC to October 22, 5:00 UTC and run the animation. Chesapeake Bay is prominent, as is the Delmarva peninsula and Delaware Bay. Orient yourself using You can freeze frame the animation.

    Note that just north of Maryland, the wind is blowing clouds from a roughly WSW direction, and that new clouds are forming over the north Chesapeake Bay. It’s fairly cloudy by 05:15 which is four hours after the photographs were taken. So while it might have been “clear” at the time the photo was taken, a small volume of slightly cooler air could form a solitary cloud before the majority.

    If you’re interested, you might look up the type of cloud that is forming, figure out its altitude, and determine the wind speed and direction from trigonometry.

  • Jonathan

    Carl, I see that there were clouds in the sky from the satellite, but the images in the photo just don’t look very cloud-like to me. I agree that there can be some very strange cloud effects, but the speed, shape and colour of the image just don’t add up to me.

    Henry, as the images were taken a few seconds apart, if it is an object, rather than an optical artifact, it was moving fairly quickly across the sky. Did you see it moving at this speed or was the image you saw somewhat more stationary?

  • Charon

    “And that light can be reddish orange.”

    Or, as the FAQ in the student orientation booklet at the University of Chicago put it, “Why is the sky orange at night?” For those students moving there from less light-polluted areas…

    Then I became an astronomer, and now quite like sodium lamps, and hate LED streetlights, because they contaminate our entire goddam spectrum, not just a couple narrow emission lines.

  • Carl Brannen

    Socorro NM had the opposite problem from Chicago. It had (and probably still has) amazingly clear skies when I was a grad student there. One summer an astrophysics postdoc specializing in radio observations (who grew up in NY city and got his PhD at Baltimore) came out to work at the nearby VLA. In conversation I found out that he was unaware that the milky way was visible to the naked eye. So I got him up one moonless morning at 4AM. We went out into the empty desert mountains and saw the milky way run from horizon to horizon. The galaxy’s light was so bright that the constellations were difficult to discern.

  • Darius

    Case closed #26.

  • Gerasimos

    The phenomenon of double or even multiple images of the sun and moon has been recorded in the past and is real. You can find a description along with some literature references of observations in M.G.J. Minnaert’s book “Light and Color in the Outdoors” (paragraph 46), latest edition by Springer. It is related to abnormal, sideways refraction.

  • DavidDeal

    there was a news report and an article in scientific american that claimed a probe (possibly voyager 1) had captured an image of a reddish sphere apparently approaching our solar system. however within two weeks the whole incident was reported as a lens anomaly. this possible object was not near any known planet or natural satellite. it was in intragalactic space. could this be the same object?

  • Chimpazilla

    I took the exact same photos from Oct 21-Oct 25th. Here is a time-lapse video I did of some of the photos, taken about 10 seconds apart, camera on tripod. Photos taken in Phoenix, AZ under clear skies.


Discover's Newsletter

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

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar