jump to navigation

A small google earth enigma July 24, 2007

Posted by dorigo in computers, games, internet, travel.

I was lazily playing with Google Earth this morning, when I discovered something weird. If you go to the Frankfurt airport vicinities, at coordinates 50°03’04″N, 8°36’43” E, you will see a Lufthansa A340 airbus taking off eastward, as in the picture below.

As you can see, there are three separate images of the plane, separated by about 630 meters. That means that the pictures composing the frame have been taken at 6 second intervals, if we assume a take-off speed of about 360 km/h (I haven’t checked, but it is the right order of magnitude). Now, the first riddle is about the shadows. If one looks closely, one understands that the pictures were taken at about noon, with the sun projecting shadows of buildings, cars, and road signs northward – at least, that is what happens to all objects in the picture (N is on the left in the picture above), save the plane! The A340 appears to project shadows westwards instead. How is it possible ?

Another riddle concerns the cars on the highway. It appears that there are only two images for them. For instance, look at the following four vehicles in the picture below: a red car, a white van, a short yellow truck, and a long white truck.

I have looked on all the highways nearby, and I could only find pairs of vehicles – but the planes are three. Maybe I did not look close enough ? What is your explanation ?



1. Tom White - July 24, 2007

the plane is casting a shadow NE, as are the buildings, trees etc in the image. At the start of the take-off sequence there are two shadow images without corresponding planes and the final image of the plane has no corresponding shadow. I can only explain these ‘phantoms’ as being producing by the method used in compositing the satellite data. As for the cars, well I have to go to work so I’ll leave that one untouched!

2. David N - July 24, 2007

For the record, those are 747s, not A340s.

3. dorigo - July 24, 2007

Darn David, you are right, it is *A* 747.

Tom, the answer is correct… And it also partly answers the second question.


4. David ross - July 24, 2007

As regards planes, cars and shadows: Since the plane is moving faster than the cars, and apparently in the direction of the photo tracking, it appears more often in the accumulated images.

However, as regards the shadows, note that the plane shadow farthest left in the images corresponds to no plane at all.

David Ross
Palo Alto

(pasted here from a different post – T)

5. dorigo - July 24, 2007

David r., that is my understanding too. As for the shadows, it is two of them, not one, that have no real plane to match, if one bases the plane-shadow correspondence on the direction of other shadows. I think the mosaic has a complex boundary in the region where the plane and shadow images sit.


6. Sam Higgins - July 24, 2007

It seems to me that the plane did not, in fact, just take off, but was flying at a high altitude when the picture was taken. (Compare its size to the size of those planes on the ground). With this in mind it seems reasonable that the placement of the shadows is due to camera not facing strait down, but instead tilted ever so slightly so that it is facing slightly to the N of E.
This is proved by examining how the tail fins of the planes on the ground do not line up with the very end of the planes they are connected to. (I am assuming that the tail fins are perpendicular to the ground)

This explains the first three shadows. However, there appears to be, somewhat obscured by the trees, a 4th, eastern most shadow in line with the first 3. . .. perhaps the last plane’s last image didn’t make the cut. . .

7. Chris - July 24, 2007

The pictures did not seem to load for me…

8. dorigo - July 25, 2007

Hi Sam,

I also considered the possibility you mention, but the plane seems so well lined up with the take off runway that it appears a bit far-fetched to claim it is just waiting for a landing signal. I do see some slant in the shadows, and I did in fact see the ghost fourth shadow (cut away from the clip I posted above). But the explanation given above by David seems more economical to me.

To solve the puzzle conclusively, I think we need to know how high above ground the image has been taken. And use some trigonometry…

However there is one further thing to note. If the plane was very high above ground, say 3-4 km, its speed would be in the 800km/h range, not in the 360 or so. Now, if you look at the distance between two subsequent positions of the cars in the highway, you gather they are separated by 150-200 meters. Now, this is consistent with 6 second intervals between frames, if the speed is 100km/h or so.

Let me actually be more accurate. The two images of the long white truck are separated by exactly 160 meters, and I found another pair of trucks very close by which are also separated by 161 meters.
Let’s assume a speed of 100+-20 km/h for the trucks. That makes
the time between images 0.160*36 = 5.76 +-1.15 seconds.

Now, since the plane images are 631+-10m and 643+-10m apart, respectively, that makes the plane speed 110+-25 m/s, that is much closer to the speed of a taking off 747 than a hovering one.

Also note that the speed appears to increase by getting farther from the runway.

Further, note that the four shadows are 604, 613, 645 meters apart as they get farther from the runway. The plane is clearly accelerating.

Finally, note that plane and cast shadow are 376 and 403 meters away tip to tip, as the plane rises above ground. That is in the hypothesis that the shadows are cast north-north-east, as all others on the ground.

And: have a look at the shadow on the tail of the planes! They line up with everything else…


9. Guess Who - July 25, 2007

TD, did you come up with this puzzle as a way to explain event analysis to unsuspecting lay audiences or is it just a case of professional deformation?

Maybe it works in reverse too: aerial and satellite intelligence analysis could be the next big job market for experimental particle physicists.

Oh, and the pictures are not loading for me either. 😦

10. dorigo - July 25, 2007

Hi GW, Chris,

I do not understand why not… Try loading them separately at
http://www.pd.infn.it/~dorigo/frankfurt1.jpg and
http://www.pd.infn.it/~dorigo/frankfurt2.jpg .


11. Guess Who - July 25, 2007

Yes, the direct links to the images work fine. There may be a browser-dependent time-out or image size limit when they are embedded in a page. I get empty frames in IE 6 and 7. Firefox displays empty frames while trying to load, then seems to give up and removes the frames too.

12. Andrew - July 25, 2007

I really like dorigo’s analysis in the #8 comment, but I think there is much simpler evidence that the 747 in question is in fact taking off during this photo sequence – the evidence comes from the growing wingspan of the plane as it moves to the east. If you zoom in all the way and measure the different images of the plane, there’s no doubt that the plane is rising.

What would be good is a reference plane for measurement – a 747 that we know is at cruising altitude. By comparing the size of a cruising 747 with one on the ground (such as the Thai airways plane parked at a gate) we could even calculate the camera’s altitude. Unfortunately, finding a 747 flying over an ocean on google earth is not such an easy task.



13. D R Lunsford - July 25, 2007

This is simple. The first shadow does not correspond to the first of the airplane images, and the last airplane images does not have a corresponding shadow recorded. If you use Google Earth’s perspective tools you can see how it all works out.


14. D R Lunsford - July 25, 2007

This is simple. The first shadow does not correspond to the first of the airplane images, and the last airplane image does not have a corresponding shadow recorded. So the first/second airplane is making the second/third shadow. If you use Google Earth’s perspective tools you can see how it all works out.


15. D R Lunsford - July 25, 2007
16. D R Lunsford - July 25, 2007

Explanation – the banding technique used in the photography did not capture the first shadow’s airplane (it was out of frame left) nor the third airplane’s shadow (it was out of frame right).


17. M K - July 26, 2007

Its easy. The plane has undergone quantum entanglement and therefore decohered into three images. Of course using Heisenbergs uncertainty principle it is impossible to say which ground state the plane was really in at the time. 🙂

18. A Google Earth enigma « Perfectly Reasonable Deviations - July 26, 2007

[…] Google Earth enigma A couple of days ago, Tommaso Dorigo wrote an intriguing post: A small google earth enigma. Here’s an excerpt of his […]

19. dorigo - July 27, 2007

Hi all,

Andrew: good point about the wing span. It is conclusive evidence. However, I disagree that finding another flying 747 would allow us to estimate the camera altitude… That datum, I think, varied during the many surveys.

DRLumsford, I think your explanation is correct and concise, but I prefer MK’s one…


20. dorigo - July 27, 2007

Oh, just a point DRL: it is two shadows that have no real plane image counterpart, not one.


21. jacen - July 27, 2007

wait a second.
if you match the first plane to the last shadow, how can you explain this:


zoomed out:

sure it looks consistent when you disreguard airplain self-shadows:

zoomed out:

still … the airplane shadows irritate me, because they are parallel to the shadows on the highway and the airport buildings.
manual forging?

22. dorigo - July 27, 2007

Hi Jacen,

no, no manual forging. B747 have a slanted tail, which will give a shadow whose edge is tilted with respect to that of vertical objects.
I actually think it is quite consistent with the projected shadows on the grouind.


23. Mauro Da Lio - July 29, 2007

It is a bug in the rendering software of Matrix.

24. dorigo - July 29, 2007

Hi Mauro,

I think you are ready for backward causation… I recently reviewed a quite serious paper on influences from the future to the past (see here, scroll down a few days). Maybe it fits with your vision of the world 😉


25. Mauro Da Lio - July 29, 2007

You men that the “real” future influences the past to get really “real” ?

26. dorigo - July 29, 2007

I don’t, they do! Actually worse: the future influences the past to prevent embarassing things from happening, like the production of too many higgs bosons… Worth a read.


27. Clay Tarver - August 8, 2007

Interesting thread. See also N W (near Auburn, New York)

Sorry comments are closed for this entry

%d bloggers like this: