Five quid we'll find Starbucks there

August 25, 2006 12:54 PM

Why are the north and south poles invisible in Google Earth?

Just now, when I just tried to look in Google Earth, I got something that more resembled Euclid's migraines than a snowy/watery wasteland.

I understand why places like Baghdad, Pyongyang, Kashmir and my flat in south-west London are photographed in such loving detail. But are there really no satellite photographs of either pole?

Or is Google Earth just unable to handle them?

(Full disclosure: I stopped studying geography when I was 13)

posted by randomination to grab bag (26 comments total)

Google maps uses a projection where the poles are pretty much unrenderable. I mean, think about the equator - tens out thousands of kilometers compressed across the 1000 odd pixels of your screen. Easy. But at the poles, as we approach the top of the sphere, we're only talking kilometers, metres of circumference stretched out across the rectangular area. I mean, imagine you're standing 10m from the north pole, and you walk East. You will be walking around in a circle with a circumference of 60m. On the other hand, walk east at the equator and you've got thousands of miles to work. And that's the problem, fitting a sphere in to a rectangle results in huge distortions - completely different scales at the equator and pole, that is somehow supposed to be expressed on the same rectangular map!

I, mean, look here - you can see the pixellation as the sattelite image is being streched out. I think Google maps pretty much gives up at 85 degrees north or south, because beyond that the image is just horizontally stretched out all proportion.
posted by Jimbob at 1:09 PM on August 25, 2006

Okay, I have some names to the projection now. Google maps uses the Mercator projection. Here's an example of that, compared to the Miller projection. Notice how even in these GIFs, the poles are cut off in the Mercator projection? It just ain't that great for showing the poles.
posted by Jimbob at 1:18 PM on August 25, 2006

(Here's another page on map projections, by the way.)
posted by Jimbob at 1:19 PM on August 25, 2006

No clue why Google Earth wouldn't have it... but both poles have most certainly been imaged.

Most earth observation satellites are actually in what's called a "Polar Orbit", meaning they would pass over both poles every 90 minutes or so. (far, far more often than they would pass over any other point on the globe).

Many of the higher-resolution city overlays on google earth are photographs from conventional airplane flyovers. But Landsat 7 can resolve visual-band down to 5m, if I remember correctly... more than good enough to see the dome at the South Pole's Amundsen-Scott base. Ikonos-2 can resolve sub-1m... good enough to show the structures in some detail.

Short answer... the photos are out there, somewhere.

(stopped studying geography when I was 22).
posted by cadastral at 1:19 PM on August 25, 2006

I'd guess that there isn't much to see there - just ice and snow. There are probably some scientific surveys, but why would a high-res satellite imaging company waste their time on the poles?

The distortion thing isn't the problem. Google maps is different from Google Earth, which uses a spherical projection. There should be no problem rendering it in Google Earth.
posted by chrisamiller at 1:21 PM on August 25, 2006

Okay, I have some names to the projection now. Google maps uses the Mercator projection.

Not to be too pedantic, but Google Earth uses a Simple Cylindrical Projection (also known as Plate Carree). Being a cylindrical projection (simplified version: imagine the earth as a sphere held within a cylinder -- drawing lines from the center of the earth out through the surface to the cylinder gives you your projection) this will not show the poles, much like Mercator.
posted by j.edwards at 1:23 PM on August 25, 2006

Oh dear. I have a continuous problem reading "Google Earth", but then thinking it as "Google Maps". You can completely ignore my answers then. Or read them anyway.
posted by Jimbob at 1:24 PM on August 25, 2006

Only tangential to your question... but here's a high-angle oblique air-photo of the South Pole, and Amundsen Scott.
posted by cadastral at 1:27 PM on August 25, 2006

Not to be too pedantic, but Google Earth uses a Simple Cylindrical Projection

Thanks for the tip - Google follows bad practice and doesn't say what projection it uses on the map, so I had to take a guess!
posted by Jimbob at 1:29 PM on August 25, 2006

I'd guess that there isn't much to see there - just ice and snow. There are probably some scientific surveys, but why would a high-res satellite imaging company waste their time on the poles?

Well... an almost insignificant fraction of satellite imagery has come from the private sector. Dwarfed by several orders of magnatude by imagery collected by world governments for physical/geographic sciences and the military.

Add to this the fact that almost every one of these satellites pass over one pole or the other every 45 minutes or so. Also, there's really no wasted time in finding some downtime where bandwidth isn't at a premium, charging up the CCD and firing off a few (hundered-thousand) shots.

End result... 100% certainty that these images exist. I'll go farther and say that both the north and south pole have been imaged down to the finest level of detail that the most capable earth-observation satellites are capable of.
posted by cadastral at 1:39 PM on August 25, 2006

Might NASA Whirlwind have it?
posted by hodyoaten at 2:10 PM on August 25, 2006

Oops... sorry, NASA World Wind. BTW here's an interesting tech article I stumbled across last week about the onboard navigation computers on Boeing jets... it appears they are not completely reliable when reaching the pole and can result in quirky behavior.
posted by hodyoaten at 2:14 PM on August 25, 2006

Low-altitude photo-surveillance satellites are actually in Sun-synchronous near-polar orbits (PDF, scroll down to page 15) which miss the poles by a few degrees but precess around the planet once a year, so that they pass over each point on earth at the same time once each day (and once each wasted night!) all year round and see consistent shadowing from earth-based features, which is important so that you can easily compare images from different passes. Therefore there is a circle around each pole over which they never pass directly.
posted by nicwolff at 2:40 PM on August 25, 2006

(Um, scroll down to page 15 of the PDF, which is page 44 of the document.)
posted by nicwolff at 2:42 PM on August 25, 2006

Rereading that, it'd be clearer if I'd italicized "at the same time". The point is that low-orbit satellites in perfectly polar orbits would fly around the planet every 90 minutes, and as the world turned under them would see the whole planet once a day. But as the earth orbits the sun, that "once a day" would happen at a slightly different time each day, and the shadows on the ground underneath would be different. So it'd be hard to tell what else had changed from one overflight to the next.

Now, the Earth isn't a perfect sphere, and it isn't uniformly dense, so its center of mass isn't right at the center of its volume. We've measured this precisely, and we know where to put a satellite so that the plane of its orbit moves around the planet at a selected rate. We put our photo-surveillance satellites in orbits that are as close to polar as possible, so that they'll see almost the whole planet, but move 1/365.4th of the way around the planet each day - which means they always have the same orientation with regard not to their initial inertial frame but with regard to the Sun.

So as the Earth turns under such a satellite, the Sun illuminates it from a constant angle, and each place the satellite images looks the same every day, all year round - unless something has moved.
posted by nicwolff at 3:42 PM on August 25, 2006

Oy, I mean 1/365.25th - what, 1/4 != .4?
posted by nicwolff at 3:45 PM on August 25, 2006

Thanks everyone, especially about the projections; that'd go some way to explaining the funny triangles Google seems to think live on the poles.

Special thanks for the photo of the research base. It seems to imply that the south pole is moving as it marked the location of the pole as it was 21 years ago.

So is there a reason why Google Earth can't just stitch the polar satellite photos onto the top and the bottom of their projections? The join wouldn't be in a place many people would care about, would it?
posted by randomination at 3:54 PM on August 25, 2006

I'm guessing that JimBob has got at least part of the story right, based on this shot from Google Maps that's way the hell North and is right in the middle of a zone where Google's losing land details but doing their best anyway.

Or something.

Really, anybody have some idea what's going on there? If you drag the map down, you see terrain. If you drag it up, you see white haze. It's an amazing shot, but I don't know what's going on there.
posted by SlyBevel at 4:12 PM on August 25, 2006 [1 favorite]

randomination, projections aren't involved here - Google Earth/Maps is using satellite photos, not scanning a big flat printed map! And there are no "polar satellite photos" to stitch in, because there are no polar satellites taking photos, for the reasons I gave above.
posted by nicwolff at 4:13 PM on August 25, 2006

Oops, yes nicwolff, I understood--but I was also referring to cadastral's first comment saying that it was almost certain the photos were out there. I shouldn't have mentioned satellites in my other comment; but if they can't be taken from satellite then they'd be taken from planes, right?

And if projections aren't involved then we lose the explanation for the blurring and haze.
posted by randomination at 4:22 PM on August 25, 2006

randomination, projections aren't involved here - Google Earth/Maps is using satellite photos, not scanning a big flat printed map! And there are no "polar satellite photos" to stitch in, because there are no polar satellites taking photos, for the reasons I gave above.

Your first statement might be wrong... your second statement is definitely wrong.

I'm not at all sure, based on what I'm seeing, that Google Earth is not remapping some sort of projection back onto a globe. Again... not saying that this is the case, but the visual depiction is kind of sketchy.

Regarding the "no polar satellite photos" assertion. You are not taking into account two things.

1. Even assuming that the orbit is "near polar"... the orbit will, at some time, take the satellite over the pole. (the pole is still, after all, "a point on the globe")

2. You are discounting the fact that most modern low-orbit RS satellites have "off nadir viewing capability" (I know for a fact that Quickbird, the latest SPoT, and IKONOS at the very least have it... and several more) which allows imagery at angles of up to 6% (or something... all from memory... so don't rely on this figure) off of "plumb". Reducing the latency period of imaging a specific location. Your earlier assertion that it passes over each spot at least twice (once?) a day is just false. with a 90'ish orbit (16ish orbits/day) there's just no possible way that it will be directly over each position of the earth with anything approaching that frequency... (which was the drive behind developing the off-nadir capabilities).

I stand firm that there exists, somewhere, sub-1m imagery of both poles only because there has to be. I hope someone more expert than I (and perhaps even with access to the images themselves) can step in and confirm.

(regardless... we've come a long way from the first recon satellites that actually ejected a cannister of film which parachuted to earth, and was collected by huge arrays of airplanes towing cargo nets to ensnare the film capsule. Blows my mind).
posted by cadastral at 4:57 PM on August 25, 2006

More photos if you need them: USAP does aerial photos of the South pole every year, from airplanes.

Also, Kite Aerial Photography (KAP) from one of the research scientists down there. The dark sector is about 1KM from the pole.
posted by culberjo at 5:29 PM on August 25, 2006 [1 favorite]

cadastral:

1. No. Picture the satellite's orbital plane transecting the Earth's center of mass and the satellite's ground track following the circle where that plane intersects the Earth's surface. Now spin the Earth. The satellite's ground path covers the whole Earth except for a circle around each pole, whose size depends on the inclination of the orbit with respect to the equator. Look at figures 5.1 and 5.2 in the PDF I linked to to above and it'll be clear.

2. You're right, no one low-Earth-orbit photo-recon satellite images the whole earth, that's why there are more than one in each orbit. I was simplifying a bit. But my point stands, if they were in polar orbits (like the radar satellites are) they'd see different shadows every season and be kind of useless. Sun-synchronous orbits at low-Earth altitude are about 8º off polar, which is about 560 miles - close LEO is only 125 miles up!

Which is not to say that I know for a fact that there aren't photo-recon satellites in perfectly polar orbits. Just means there very well might not be.
posted by nicwolff at 5:50 PM on August 25, 2006

Nicwolff is most assuredly in the wrong about satellite images and projections. While not all images are projected, (indeed when you obtain raw imagery it is 'raw') any two images that 'line up' to form a seamless image must be projected. Simply viewing one satellite image next to another is all the proof you need that some kind of mathematical distortion has been applied to account for the curved surface of the Earth appearing flat (in the sat image) on your monitor.

Randomination said: "It seems to imply that the south pole is moving as it marked the location of the pole as it was 21 years ago."

The poles are constantly moving. See this article and the associate figure which shows the progress the North pole has made since 1600. Also note the difference between magnetic north and true north.
posted by maxpower at 8:04 PM on August 25, 2006

Individual hi-res photos are being mapped back onto the terrain, maxpower, and that might produce some slight local distortion - but that's got nothing to do with this question, it'd work just fine at the poles.
posted by nicwolff at 10:21 PM on August 25, 2006

Really, anybody have some idea what's going on there? If you drag the map down, you see terrain. If you drag it up, you see white haze. It's an amazing shot, but I don't know what's going on there.

WAG, but could this be the result of the Aurora Borealis?
posted by scrump at 10:03 AM on August 28, 2006