Followup: City lights from space

By Phil Plait | May 18, 2011 7:00 am

Last week, I posted a picture taken by International Space Station astronaut Ron Garan, showing city lights at night. I was curious about what the cities were, and initially had a hard time figuring it out. But, using various tools — Wolfram Alpha to get the ISS position at the time the photo was taken, an atlas, and a rough direction and scale using the visibility of Orion’s belt in the picture — I took a stab at which cities were which.

A followup on that post is in order. A lot of people left comments on the post, and some went through similar sleuthing exercises as I did. Quite a few agree with my assessment: the city in the center of the picture is Warsaw, Poland. Berlin is on the horizon, and other cities are as I determined.

I used an old-fashioned atlas to find the cities because I couldn’t find a good piece of software to do it. But then I had a facepalmy moment when several people suggested Google Earth! I got a new computer when my Mac died a little while ago, and never installed GE on it. So I downloaded and started fiddling with it, and sure enough the cities do seem to match. If you’re curious, go to latitude 50.56 North, longitude 38.23 East, zoom out to an eye altitude of about 221 miles, and rotate the picture so you’re facing west.

At this point, you’ll see a big city centered in the frame that turns out to be Kiev, Ukraine. That threw me for a second, but I remembered that given Orion’s size in the ISS photo in question, it’s clear the picture is zoomed in a bit. Kiev is off the bottom of the frame. Also, in Google Earth, you can overlay the landscape with a NASA picture of cities at night (in Layers, go to Gallery then NASA then Cities at Night). Zoom in so that Warsaw is centered, and the other cities do indeed appear to line up with my initial guess.

So hey, I was right! Richard Drumm, an astronomer who lives in my old stomping grounds in Virginia, kindly annotated an image to show what’s what. You can find more in the comments of the original post.

And I’m honored to have discovered that Ron Garan, the astronaut who took the picture, tweeted about my post as well!

All in all, this was a fun and interesting exercise. It gave me an appreciation for what cartographers do, and also how hard it must be for astronauts to figure out where they are over the Earth. Imagine waking up, looking outside your window, and having no clue at all even what part of the Earth you’re seeing! Before teh internetz, I used to spend a lot of time trying to do something like this when I was identifying stars in photographs. When I first looked at Ron’s picture I figured it would be easier looking down, but I was wrong.

It also was rewarding in a more literal sense, in that I got to find Google Earth all over again, and see that Wolfram Alpha is a pretty useful site. Next time I have a weird question, I’ll have to try there first.

And, of course, the most important lesson here is that playing is fun, even when it’s nerdy and maybe seems at first like a time sink. In a lot of cases, playing leads to learning, and there’s always room for that… and there’s always much, much more to learn.


Comments (14)

  1. Thanks for the shout-out, Phil!
    Yeah, I love puzzles like that. I just used the Maps app on the iPad and looked at the angles and alignments of the cities to figure it out. I started with mentally drawing a line from Berlin to Warsaw and found that Poznan was in the right place.
    Then Lodz dropped into place on the line from Warsaw to Dresden.
    Great fun!
    Hardest part was composing the text & arrows in PhotoShop, believe it or not! I had to drop the font size a bit to help things fit over by Krakow.

  2. Sam H

    @Phil Plait: Imaging waking up, looking outside your window, and having no clue at all even what part of the Earth you’re seeing!

    Indeed – in GE I once switched clouds on and everything else off, went down to a roughly ISS-ish altitude, rotated the view to oblique and started scrolling. It was liking I was scrolling over an alien landscape – most landmarks I couldn’t recognize from such an unfamiliar angle. Of course, the ISS crew is much more skilled in the respective disciplines for finding their way around, but let’s say they were in the South Pacific in a region without islands; nothing to see but water. Would they still be able to roughly approximate their position based on things like the height, angle, trajectory etc. of the sun, moon, and stars as they orbit?

  3. Kyle

    @Sam H: Sure they could just have to do a little sextant work and they could figure where they were because it’ll take them a few minutes to do the maths and look up the appropriate tables. Just good old basic celestial navigation.

  4. Kyle … I wonder if the ISS is orbiting too fast for sextant-based celestial navigation?

    In fact, can you shoot the sun/stars from low-Earth orbit; does that work?

  5. Mark

    An astronaut friend of mine who flew with Story Musgrave said that Story could recognize most oceans and bodies of water by their _color_ and hue.

  6. ‘New computer’? (Not a new Mac?)

  7. MartinM

    Imagine waking up, looking outside your window, and having no clue at all even what part of the Earth you’re seeing!

    I’ve had mornings like that.

  8. JohnDoe

    Before the internetz, you couldn’t just upload a random picture of the sky and have it automatically identified and annotated, even if it was mirrored, scratched and vignetted, as now does at

  9. Ross

    One of the necessary skills for an airplane pilot is the ability to match up what you see out the window to the map. The effect is basically the same, looking at the ground from an airplane at 6,000 feet is like looking at a map at a very shallow angle, 5-10 degrees. Shapes are distorted and more difficult to identify (the curve of a road, the outline of a lake), so I tend to look for intersections, edges, and discrete landmarks such as stadiums, bridges, water towers, and airports of course.

    At night, the view is remarkably similar to the posted photo, just a different scale. The small dots of light would be individual street lights, and the large patch might be a shopping mall. The yellow areas on the aeronautical charts are supposed to depict the actual lighted areas at night (see

  10. Mark

    I love wolfram alpha! It’s especially helpful for differential equations.

  11. Physicsman

    I think another lesson after having tried Google Earth and seeing that numerous people in the thread beat me to it is that when you have a pool of educated followers cloud sourcing them can produce amazing results. I think in your next blog you should post some random equation and state that it is the ‘newly’ discovered equation unifying the four known forces but you not sure about the order of the terms.

  12. Don

    If you like differential equations check out this video. It’s a prototype of an interface on an iPad that allows you to directly manipulate parameters and coefficients to generate plots. Very cool.

    In order to bypass moderation, go to vimeo and look at video number 23839605. Just put the number after the slash in the dot com with a slash following. If you are searching for it, the title is “Interactive Exploration of a Dynamical System”.

  13. Caleb Jones

    I have a genuine astronomy question.

    I understand that one of the techniques to detect exoplanets is to measure the delta in luminescence of a star that occurs when a planet transits in front of it.

    Is it possible to, through some other independent techniques, figure out the size of the planet? If so, could that data be compared against the luminescence delta to see if there is a significant mismatch? And if so, could that be used as a way to indicate that the exoplanet itself is emitting light which might indicate intelligent life and technology?

    There are a few steps to this and lots of room for problems from the lack of precision in our instruments, but it seems like it might be possible.

    Anyone have any insight on this?

  14. Physicsman

    Caleb Jones:

    Disclaimer: I haven’t checked any of this to see if its reasonable!

    My guess is, you would need the precise mass and size of the host star and the precise period of brightness change to establish an orbital radius and velocity. Using the radius, velocity and the precise rise (dim) time from max bright to least bright you could establish an estimate for the diameter of the planet. Knowing this you could predicted the amount of dimming you would expect and compare it to observed. I would be amazed if it is even remotely possible for the nearest known exoplanet. I think your best bet would be, after having found an exoplanet with a known period, to point a radio telescope at it (much better sensitivity and resolution then optical) and see if you pick up a late night monologue in alienese.

    Your best bet is SETI. That is if it was still on…


Discover's Newsletter

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


See More

Collapse bottom bar