31
$\begingroup$

In this satellite image of Madrid airport (MAD/LEMD)...

Screen cap from Google maps
Source:Google maps satellite view
Alternate link from maps.google.nl still renders incorrectly for me, but appears fine for others...

...it appears that there is a very noticeable left kink in the landing area of runway 32/33R.

According to what I can interpret from Jeppesen's airport data, there doesn't appear to be any indication of this kink:

Jeppesen airport diagram
Page 50 of Jeppesen hosted by uvairlines.com.

Is that kink actually there, or is it an artifact of Google's processing of different images? There is obviously something going on, as you can see both RWY32 & RWY33 markings in the image, while Jeppesen clearly identifies it as 33.

If that kink is actually there, what impact does that have on a pilot landing? I believe the touchdown target is the two long, solid bars at the top left of the image, which excludes the kink, however, there is enough rubber to indicate that touchdowns have occurred much closer to the threshold than that.

$\endgroup$
  • 2
    $\begingroup$ Interestingly I just followed your link and I see the kink too with the 32/33 makings unlike the clear 32 in DeltaLima's answer. $\endgroup$ – Notts90 Jan 18 '17 at 15:40
  • 1
    $\begingroup$ Glad to know I'm not totally crazy and/or being singled out by Google for mind-control experiments, @Notts90 ! :) $\endgroup$ – FreeMan Jan 18 '17 at 15:41
  • 3
    $\begingroup$ Maybe they double runway numbers are to account for the kink. The first part of the runway is 33R and the after the kink is 32. :) $\endgroup$ – reirab Jan 18 '17 at 19:52
  • 2
    $\begingroup$ My guess is you have got images of two image sets, one at least being not accurately geo-referenced.. The new set is copyrighted 2017. It's clear two images overlap. "33" is part of the old set, and is the top image, the bottom image ("32") is from the new set. The set used may vary with zoom level. $\endgroup$ – mins Jan 19 '17 at 8:11
  • 1
    $\begingroup$ @mins I understand the copyright date on Google Maps / Google Earth is always the current year. I think only Google Earth shows the actual date of photography. $\endgroup$ – Dai Jan 19 '17 at 18:03
59
$\begingroup$

An interesting feature of your question is the apparent double-exposure of runway number in your image. And I suspect it will also account for/explain why you see a kink in an image.

Occasionally, runways change number. The reason for this is basically that runways only have 36 possible designations (1-36) however a runway designated 18 may be orientated anywhere from 175-184 degrees. Usually the actual orientation of the runway is rounded to the nearest 10. hence, a runway pointing at 184 degrees would be designated "Runway 18".

Magnetic variation drifts over the years (eg, 10 years ago Southern UK was +3 degrees, now it is +1). This may make a runway which was once 184 degrees magnetic now 186 degrees and that would require its designation to change from "Runway 18" to "Runway 19".

Now, as for your image, I would guess that Google have taken a grab when it was Runway 32, and (at least partially) again once it had been re-designated (and repainted!) Runway 33. This is backed up with the airport's wikipedia article:

On 20 September 2012, both runways 15/33 were renamed as 14R/32L (the longest) and 14L/32R (the shortest).

The act of stitching images together, and the image processing algorithms in use, mean you get kind of a partial of one image and partial of the other. It is almost certainly not there in real life, and thus has no effect on the pilot's handling around that runway.

$\endgroup$
  • $\begingroup$ Actually, @DeltaLima did mention the runway renumbering in his answer, though he didn't provide a reference. $\endgroup$ – FreeMan Jan 18 '17 at 16:32
  • 2
    $\begingroup$ @FreeMan I felt he made a very passing comment to it, and didnt really address that that was clearly the source of your distorted image. I thought you might like some background as to why runways might change numbers. hence the answer. $\endgroup$ – Jamiec Jan 18 '17 at 16:38
  • $\begingroup$ Fair enough explanation. $\endgroup$ – FreeMan Jan 18 '17 at 16:41
  • 3
    $\begingroup$ @FreeMan Also the 3 answers you have are "Works on my computer", "Works on my computer" and "Its just the image". Nothing I'd call a "good answer to the question" - with no disrespect to other answerers. $\endgroup$ – Jamiec Jan 18 '17 at 16:43
  • 1
    $\begingroup$ This sort of kink appears very frequently throughout Google satellite photos. $\endgroup$ – Michael Hampton Jan 19 '17 at 1:01
31
$\begingroup$

Did you bend your screen?

My google maps is better behaved:

enter image description here

Also note that the 15/33 runways were renumbered to 14/32 in somewhere around 2012 due to change in the magnetic heading.

$\endgroup$
  • $\begingroup$ Interesting. I just realized that the date on the Jeppesen plates I found was 2008 and had it marked as 33. Interestingly, even a Ctrl-F5 (forcing Firefox to refresh the image & not pull it from the cache) is still showing a bend. Though, when I zoom out a bit, it's less of a kink from, say 325° to 321°, but more of a chicane to go around that white circle the to left of the 2nd set of 6 bars. Very interesting, and you'll likely get the check for posting a Google map image vs ROIMason's Bing image - I'll give it a bit to see how many other map images show up. :) $\endgroup$ – FreeMan Jan 18 '17 at 15:27
  • $\begingroup$ Based on Notts90's comment on my OP, are you using the US version of Google maps, or are you in some other country using a localized version? i.e., can you provide the source link for your image? I wonder if there are different images on different servers... $\endgroup$ – FreeMan Jan 18 '17 at 15:43
  • $\begingroup$ I am in Germany (.de), using the Dutch (.nl) Google maps. $\endgroup$ – DeltaLima Jan 18 '17 at 16:09
  • $\begingroup$ which, interestingly, still renders incorrectly for me! It looks exactly the same as my maps.google.com link. Oh well, I'll just chalk this up to "one of those things"... $\endgroup$ – FreeMan Jan 18 '17 at 16:16
  • $\begingroup$ @DeltaLima Better even, I can see the dent in one browser (with Maps lite) but not in the other. So it's not only an image processing artefact, but also a caching one $\endgroup$ – Bergi Jan 18 '17 at 18:07
21
$\begingroup$

This image from Bing maps shows the same region, but without the kink, so I believe this is an image-processing artifact.

Find it on Bing Maps via this link

enter image description here

Source: Bing Maps

$\endgroup$
16
$\begingroup$

The question is whether the runway is straight, so let's start by finding something handy that is a known straight thing.

Notice that the wear marks on the runway (the darker centerline) also slice to one side with the runway - and the 1st law of motion would suggest that is unlikely. This is why I would bet a lot of money on it being an artifact, and not reality.

[lolz] Of course there could be a frequent sideways wind gust that always happens right there, causing the airplanes to move sideways at the exact place that the runway itself is bent...but then why stop there? Maybe the construction crews were similarly affected by that wind, and that's why they built it crooked in the first place![/lolz]

runway image showing sideways slice in wear marks

$\endgroup$
  • 1
    $\begingroup$ "Maybe the construction crews were similarly affected by that wind...". Naaah. Sangria. $\endgroup$ – Bob Jarvis - Reinstate Monica Jan 20 '17 at 1:23
  • 1
    $\begingroup$ Maybe that wind moved that part of the runway. :-) $\endgroup$ – WGroleau Jan 20 '17 at 23:18
7
$\begingroup$

This is most likely an artifact of the error correction in aerial orthophotography.

Planes performing aerial photography or LiDAR will fly a 'windrow' pattern while photographing the landscape, producing long strips of images.

Nadir courtesy of ESRI
ESRI image

This raw images are then processed and stitched together to create a composite image and orthorectified. This process is largely automated, the algorithm using similar points as reference points to stitch the images together. Similar process here, done manually. Sometimes the edges don't line up perfectly for various reasons: flight vector correction, different lenses, distortion in a lens, etc. The computer then when stitching the images together stretches or squeezes the image to fit the edge between two known, good reference points.

It interpolates the image which means that sometimes the edges between the reference points don't line up, therefore producing the weird kink in your runway. These errors are common and often found on linear objects where the break is easily recognised.


Re-reading Jamiec's answer about the runway designation: This is old photography being stitched with old photography just like the answer suggests; the technique is the same and the artifact would look similar too!

$\endgroup$
0
$\begingroup$

FWIW, the other end of 32/33 is labeled 14L. This means (I believe) it is the left of a pair of runways at heading (as you enter it) 140°. Thus the 32/33 end should be labeled 140+180 = 320°, i.e., 32, and since it's now on the right it would be 32R. I measured the actual heading, taking horizontal on my screen to be true east-west (not necessarily so, depending on the details of Google's geographic projection) and got 323°, also consistent with 32 as the correct number. Finally, the kink occurs where a distinct change in background color happens, as one sees often and which is presumably where two different images were joined. The duplicate numbers are in the overlap region. The 33 is bogus (but where it comes from I can't see at all).

$\endgroup$
  • 1
    $\begingroup$ The other answers explain why there is an image of the runway labeled as 33. $\endgroup$ – fooot Jan 19 '17 at 21:09
-1
$\begingroup$

No, this is just an image registration artifact.

$\endgroup$
  • 24
    $\begingroup$ This answer could be improved by providing either an image showing the runway is straight, or providing more info about why this artifact might be here, as the other answers do. $\endgroup$ – fooot Jan 18 '17 at 16:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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