EDIT: The stories in Phys.org NASA's EPIC view spots flashes on Earth and Live Science Mysterious Light Flashes 1st Spotted by Carl Sagan Explained seems to have information potentially relevant to this. Apparently these have been seen for decades by several different spacecraft.

Journal Article now viewable as epdf: http://onlinelibrary.wiley.com/doi/10.1002/2017GL073248/epdf

Watch the NASA Goddard video EPIC Observations of Ice in Earth's Atmosphere:

This answer has been bugging me for a while, so I would like to challenge it by asking for a technical reference, journal article, or something that explains the phenomenon more fully.

This purple dot appeared around the sub-solar point in an EPIC photo of the Earth, from the DSCOVR satellite in a Lissajou orbit around the Sun-Earth L1 point.

Is the purple dot really caused by an effect of Cassegrain optics and/or coherent backscattering? If so, can I read further on the technical explanation of small, highly collimated bright spots at sub-solar points on Earth seen from space?

Note, this looks to be of the order of 10km or less at a distance of 1,500,000km! That's what I mean by highly collimated!

Note also that it is of the order of 1 pixel of the EPIC's camera as well.

above: Bright spot at sub-solar point taken from this question.

above: Screen capture from Google maps, annotated with a box marking the approximate location and size of the cropped image with the purple dot.

above: The full DSCOVR EPIC image taken from this question. If you right click and open in a separate view, and zoom, you can see the purple spot.

Side note:

Bright, highly collimated reflections could be produced by highly reflective, fairly flat surfaces, and water is one such example.

Watching this video one can clearly see occasional bright flaring near the sub-solar point when it is on water, presumably the water is unusually calm and a much larger fraction of sunlight is directed into a cone around normal incidence.

above: Screen shot from DSCOVR video showing an example of a common bright reflection from calm seas near the sub-solar point.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – called2voyage
    Commented Feb 7, 2017 at 17:45
  • $\begingroup$ An interpretation of the answer has been proposed that - if I understand correctly - has the camera taking multiple photos to "see around" its own secondary mirror, and this plus coherent scattering from earth requires image processing which then produces the purple dot. This still does not make sense to me. There isn't any image processing you can do to "get rid" of coherent backscattering because 1) the satellite doesn't move to a different place to take the other picture, and 2) coherent scattering - if it was happening - would be a real part of the Earth's appearance. $\endgroup$
    – uhoh
    Commented Feb 7, 2017 at 18:56
  • $\begingroup$ Wait, if this is a mirror telescope - then the second mirror sits squat in the middle of the telescope tube, right in front of the center of the primary mirror. Image of Earth hits the primary mirror, gets reflected onto the second mirror and then into the camera (to the side). Except the part obscured by the second (tiny) mirror. Care to check if the dot is perfect geometrical center of the image? $\endgroup$
    – SF.
    Commented Mar 9, 2017 at 23:56
  • 1
    $\begingroup$ A different approach: The location seems to have an exceptionally high albedo comparing to surroundings; pure silicon crystals were found in the area, I couldn't find what gives the ground this coloration but pure white is not uncommon. With the satellite in L1 and this being the central point (sun in zenith at that point at that moment) plain reflection can oversaturate the camera sensors. $\endgroup$
    – SF.
    Commented May 17, 2017 at 11:33
  • 1
    $\begingroup$ photo.stackexchange.com/questions/89488/… $\endgroup$
    – SF.
    Commented May 19, 2017 at 7:27

1 Answer 1


I think this was answered perfectly well in both the answer and the comment left by Tildal.

So it would appear as an overexposed region, that might appear as pretty much anything after post-processing, including the Moiré fringed dot with typical read and blue channel interference pattern around it on the image that you noticed, once the RGB sources, or slightly different angles, or both, are interpolated. Notice that there's also a bit more of color bleeding around the central dot, than elsewhere on the image

Moiré pattern removal tool removes color bands nicely without loss of sharpness, leaving an artifact resembling a spherical object roughly 40 km (25 mi) in diameter. It doesn't appear to be a dead pixel on the sensor, but a post-processing artifact of 5x5 pixels in the central region of a 2048x2048 pixel source

The fact it is purple is almost irrelevant - post processing of images can leave you with results of any colour.

  • $\begingroup$ Read the discussion between me and uhoh in chat. Apparently, this answer is not enough of an explanation for him. $\endgroup$
    – called2voyage
    Commented Feb 7, 2017 at 21:05
  • $\begingroup$ You've simply reposted some of the text I'm asking about as your answer, adding only your opinion that it was answered perfectly well. This is not in any way a proper stackexchange answer. Can you at least explain the role of the secondary mirror as discussed by the first paragraph in the original answer and the large graphic there, and how coherent backscattering requires image post-processing? I'm looking primarily for something technical about the phenomenon cited where I can read further, or a better technical explanation. $\endgroup$
    – uhoh
    Commented Feb 8, 2017 at 3:45
  • $\begingroup$ If coherent backscattering is indeed happening, then it is a true part of the appearance of the earth - it is real. It is not an artifact like scattered light inside the telescope, it is one of several physical processes that make up the reflected light of objects. Can you cite another example of a spacecraft telescope having to use tricks to hide coherent backscattering? $\endgroup$
    – uhoh
    Commented Feb 8, 2017 at 3:52
  • $\begingroup$ Question has been updated with new information. $\endgroup$
    – uhoh
    Commented May 16, 2017 at 9:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.