Darwin "Meteorite" Image

Darwin ImageOn November 22, 2004, an interesting series of images was made by Wayne Pryde in Darwin, NT, Australia. Mr Pryde was taking a series of digital images to show cloud buildup. One image from the series appears to show a long streak ending in a flash. It has been suggested that this image is recording a meteorite fall. The image series has been posted on the Astronomy Picture of the Day site, and discussed in great detail online. (Note that the before and after pictures on APOD are swapped.)

Was this a meteorite? Almost certainly not! (If you wish, you can skip the following analysis and proceed straight to my conclusions.)

First, some technical details. Time and exposure information was extracted from the EXIF headers in each image. The camera was a Canon PowerShot G3.

  • Image 1, 18:52:37, 1/20 second @ f/5.6, pattern metering, flash fired
  • Image 2, 18:52:52, 1/20 second @ f/5.6, pattern metering, flash fired
  • Image 3, 18:53:07, 1/20 second @ f/5.6, pattern metering, flash fired

The second image shows the effect, with the images taken 15 seconds before and after appearing normal.

Some additional details are posted in this news story. From this it is possible to determine that the image was made at 12° 28' south, 130° 50.5' east. The camera is facing a little east of south, and is about 500 meters from the wharf. At the time these images were made, the azimuth of the Sun was 249° (WSW) and the altitude (with refraction) was -0.3°, or just below the ideal horizon. However, the Sun's azimuth placed it behind land, so the actual horizon was higher.

Other observations that I make upon close examination of the image are:

  • The darker path is very straight, with little sign of an arc. An object falling ballistically always traces a parabola. In order to appear straight, the velocity would have to be extremely high. The line is also slightly wavy, but with no regular period. There may be a hint of an arc, but this is almost lost in the noise of the waviness, and is partly an illusion created by forcing the endpoint to coincide with the flash. However,
  • The line defined by the path does not intersect the flash, but rather a point 10 to 15 pixels above it.
  • The line appears to begin and end wholly within the image.
  • The flash does not appear to be located at the location of the luminaire atop the light pole, but 8 to 10 pixels to the right. If the flash were at the same distance as the lamppost, this would be nearly a meter from the luminaire.
  • There is no sign of the "smoke" from the second image in the third image, 15 seconds later.
  • There is no sign of anything in the image being directly illuminated by sunlight, except for the high clouds. This is consistent with the Sun being below the horizon.
  • With the exception of a few bad pixels, the only saturated areas found on any of the images are in the brightest part of the clouds.

The Theories

Many theories have been proposed to explain this image. I will discuss each of the more plausible ones. I will not discuss the many crank theories, such as lasers, particle beam weapons, UFOs, etc.

1. Fraud - the picture may be a hoax, produced by digital manipulation. While this can't be completely discounted, there is strong evidence that the image is genuine. All three of the images have their EXIF headers intact, and there is no sign of manipulation by any processing application. All three images have nearly identical signal-to-noise values across common areas, something that would be difficult to achieve if only one were doctored. There is also no sign of double JPEG encoding (the decimation boundaries are the same on all three, and are regular). It also doesn't look like most people would expect a meteor to look, and the point of the "explosion" doesn't correspond to any physical object. These seem like odd choices for someone perpetrating a hoax, when the other technical details were handled so well.

2. Meteorite/Space Junk - on first examination, this sounds reasonable. After all, thousands of meteorites fall to the Earth every year, and with so many cameras now in use, one might be captured. But there are too many things wrong with this theory. Normal meteorites fall at terminal velocity, from 50 to 200 meters per second. They fall nearly vertically, deviated only slightly by wind. For the last several minutes of their fall, they are cool and don't leave a smoke trail They don't explode on impact. The trail seen in the image is far from vertical. A meteorite still moving fast enough to leave a visible trail would have started out several meters across, and would have produced a brilliant fireball, sonic booms, and other effects that would have been widely observed. If this image shows a meteorite at the distance of the wharf or beyond, it would be traveling several km/s. Anything massing less than a few tons, traveling at such speeds in the lower atmosphere will experience a deceleration force of hundreds of Gs, well over the material strength of natural materials. Similar arguments argue against space junk, although it generally begins its descent somewhat slower than most meteorites.

3. Explosion of a Lightbulb - it is suggested that the light was just turning on, and exploded. Certainly, light bulbs can do that, but it is very uncommon for a high wattage gas bulb. These types of lights usually have a double glass envelope, are are mounted in fixtures with safety glass. An bulb might flare brightly when it burns out, but it won't explode and release gas or debris. In this case (and ignoring the fact that the "explosion" doesn't seen to line up very well with the location of the bulb) the light fixture has been examined, and found to be mechanically undamaged (although nonfunctional, for reasons not yet given). I don't believe that an explosion large enough to produce the effect seen in the image could occur without some visible signs of damage to the fixture. There has been some suggestion that this isn't a light at all, but the mast of a boat. I don't believe the image supports this- there are several identical posts visible, and their tops are all in exact alignment. It is possible that the white wisp seen to the right of the flash is smoke rising from a boat, and the smoke is catching the light of the lamp turning on (or burning out). I think the flash appears too bright for this, but I would not completely discount the possibility. If so, the dark streak in the sky is an unrelated phenomena, unless...

4. The Dark Streak is the Shadow of the Lamppost - this is a popular theory on the online discussion, but I think it makes little sense. A bright light just below the lamp would have to be lighting up the whole sky in order for the lamp to cast a visible shadow, yet there are no reflections on any of the structures around the pier. The image of the "explosion" is not even bright enough to saturate the CCD, and the sky brightness is constant between the three images. Furthermore, the air is obviously quite clear, given the view across the harbor. I simply don't think there are enough particulates or vapor in the air to create such a distinct shadow. Also, the shadow would need to be directed away from the camera in order for perspective to keep the width uniform, which is inconsistent with the geometry of the light fixture.

5. The Dark Streak is the Shadow of a Contrail - this is also a popular theory, and certainly the streak looks very much like the shadow of a contrail. I've seen and photographed many similar shadows. Contrail shadows are usually planar, and look very straight even when the contrail is not. But seeing them depends on the observer being near the same plane, so that a large thickness of the shadow is viewed. This happens when the contrail, Sun, and observer on all in line with one another. This is clearly not the case here, however. The Sun has just set to the right, and any light coming from it is at right angles to the camera direction. Contrail shadows are sometimes seen cast onto nearby clouds or thin vapor layers. But there are no uniform clouds in the area of the track- only scattered clouds at various distances, which would fragment any contrail shadow cast directly on them. There is simply no place in the sky that a contrail could be to cast a shadow like this.

6. The Dark Streak is a Crepuscular Ray - these (and anti-crepuscular rays) are formed by clouds or landscape features when the Sun is near the horizon, even slightly below it. But they converge on the solar or anti-solar points- both of which are outside the frame of this image.

7. The "Explosion" is a Reflection - there are two theories here: the flash is a reflection of the Sun from the top of the lamppost, or a reflection of the sunlit clouds from the water. I discount the first possibility on the grounds that the Sun has already set. If it were able to reflect from the top of the post, we would see evidence of other structures in the image that are sunlit, and none are. A reflection from the water is a possibility, and certainly the color of the flash looks very similar to the color of the cloud above it. But such a strong reflection would require a rather large planar zone in the midst of what looks like a mildly choppy bay. I can't see what would cause that.

8. A Missile, Model Rocket, Artillery Shell, etc. - something launched from the ground could easily have a supersonic velocity over the exposure time of this image. Something launched from the air towards the wharf could also have a high velocity. But this theory doesn't explain the midair flash. Small model rockets don't have such straight trajectories, and it seems very unlikely that military air-to-surface weapons would be in use in an area like this, near the Darwin airport flight pattern, or that surface-to-air weapons would be in use on a moderately secure commercial dock.

9. Some Kind of Camera Artifact - I have many years of experience with digital images, and I don't know of anything that could produce this kind of effect. All sensor defects produce artifacts aligned with the rows or columns of the sensor- never diagonal. Most sensor defects are seen in all images, while this shows up in just one, taken between two images that show no artifacts. Cosmic ray hits produce saturated pixels, in some cases small zones of saturated pixels. But the "explosion" is not saturated. Cosmic rays can produce showers of secondary particles that show as a linear structure of bright pixels, but never as a long, linear, dark zone. There is no evidence of any kind of software bug, or JPEG encoding artifact that could cause any of the anomalies seen in this image. The "explosion" is below the saturation level, so there is no reason to expect artifacts around it. Nor does anything suggest an optical artifact, since there appears to be nothing to cause a flare, and the long dark streak does not resemble any common (or uncommon) optical artifact at all. Any failure of the sensor would continue to show up on later images.

10. Lightning - a possible theory for some elements of the image, but not all. While the flash itself could show a lightning strike- either a low energy leader strike, or the result of a larger strike just before the shutter opened, it can't explain the dark streak in the sky. Some have suggested that it is the leader strike, but it seems far too straight for that. Also, I would expect the leader to either be invisible, or somewhat brighter than the sky background. How could it possibly be darker? Some other electrostatic phenomenon like ball lightning or St Elmo's fire might also explain the flash, but again, not the streak. Furthermore, the photographer reports that he observed no lightning during his imaging session.

11. Something Moving Very Near the Camera - this theory assumes that something passed in front of the lens. Possibilities include a piece of vegetation, a hair, a spider, or an insect. The last seems most likely. The "explosion" near the lamppost is simply the insect being illuminated by the camera flash. I find this the most plausible explanation, because it very simply explains all the features seen in the image. The straight dark line is the silhouette of the insect against the sky. It is about 2% less bright than the surrounding sky, which is about right for an insect that spends only a short time in any small part of the path. I'm not sure about the flash mode the camera was in- in some cases, the flash fires at the beginning of the exposure, and in some cases at the end. But that only affects which direction the insect was flying. The wispy smoke may actually be smoke from a boat or vehicle on the wharf, or it may be part of the insect being illuminated (transparent wings would not show in silhouette, but would reflect light from the flash). Given that the flight path would only be a few centimeters, its straightness is not particularly unusual. A difference image between the second and third shots seems to show that the path is slightly wider at the upper left, and varies in width slightly over its length. This would make sense if the insect is a little closer to the camera at the left, and if the profile of its body is changing as it flies.

Further supporting evidence for this theory comes from some fine work done by several participants in the online discussion at Night Sky Live. This includes an FFT analysis that may just be pulling a wing beat frequency out of the image noise along the streak, several images of flying insects that show similar effects, and an elegant experiment with a dead bee outrigger on a spinning bicycle wheel that produced a nearly identical dark trail and flash as that seen in the Darwin image. (I'd like to give credit to these scientists, but they are all using screen names, in some cases non-unique ones, and I don't really know who they are.)


My Conclusions

With near 100% certainty, the image does not show:

  • a meteorite or piece of space debris
  • a shadow from the lamppost
  • a contrail, or shadow of a contrail
  • a crepuscular or anti-crepuscular ray
  • a camera artifact

Still very unlikely, but impossible to completely discount, are:

  • fraud
  • an exploding light bulb (doesn't address the streak)
  • a reflection from the sunlit clouds (doesn't address the streak)
  • a rocket or high velocity weapon
  • lightning

Which leaves me with the most likely explanation: this is simply the image of an insect passing right in front of the lens. The theory is well supported by the image analysis and by experiment, it is plausible (a resident of Darwin reports that flying ants and other insects are numerous at this time of year), and it requires no extraordinary coincidences.


© Copyright 2004, Chris L Peterson. All rights reserved.