The bones of the earth; a new perspective

[fivemack]

The Shuttle Radar Topography Mission is one of the esoteric topics I'm most prone to enthuse about; I think of it as the last major donation made by the military-industrial complex to the sum of human knowledge before 9/11, and so possibly the last for some time. Using a fortnight of time on the Space Shuttle, Earth was mapped to 90-metre resolution in position and about one-metre in height.

This is a very interesting resolution; water erosion is the main visible feature, you can see some of the moderate-scale processes of geology (are the long linear features dykes?), and you can't see humanity at all.

ESDI is the best interface to the data I've found so far, though it dispenses data in TIFF-format files, one square degree per file and one 32-bit float per pixel; these aren't viewable in standard software.

If you pull out hexdump, gcc, python and gimp in some order, however, you can get something like



I'll be very impressed if anyone knows the shape of my country well enough to be sure of where that is.

I've got a full-scale version (5MB .PNG) or a version with labels (5.5MB .PNG); I'd recommend downloading these to anyone with the connectivity. The labels come, I think, from a full dump of the National Geospatial Intelligence Agency's GNIS database; coordinates there are rounded to the arc-minute, which is about a mile, so I've positioned them randomly within a two-arc-minute square aroud the given position. This may mean North Wallop is positioned south of South Wallop; sorry.

If anyone's aware of a better freely-available gazetteer, please tell me.

Comments

[ewx.livejournal.com]

I initially wondered about the Southampton area but skew, but then decided it was the Severn estuary. Something about the angle of it. You'll have to trust me when I say that I only looked at the labelled version after that.

Oh, and I have managed to line up two browser tabs onto the same image such that I can turn the labels on and off just by switching between tabs. Which amuses me after a couple of glasses of wine.

[jojomojo.livejournal.com]

You can probably feed the data to the Demeter engine ( http://www.terrainengine.com ) and view it in 3d, too :)

[kaet.livejournal.com]

Is it the Bristol area?

[fivemack.livejournal.com]

Yes, Bristol is in the bottom left (Bath is in the middle of the big black star of rivers, and Bristol is in the flat patch where they flow into the sea). The big bright feature in the centre is the escarpment which towers to the east of Cheltenham; the little hills, which would be pretty off-shore islands full of seagulls if the sea rose high enough, make that region in this display remind me strongly of the shores of Xanadu on Titan.

[kaet.livejournal.com]

Fortunately I have CSI magic computer software (aka GIMP) and you can find the two Severn Bridges (and some other linear features) from the big image, :).

[ewx.livejournal.com]

It was the bridges that decided it for me.

[aldabra]

Hi. Have you got a nice easy-to-use sea-level-rise mapper? Because I can't find one. Where I could type in a sea level rise of 57m and see what's left standing? Ideally at street map level...

I don't have the software to distribute, but I've got the capability

[fivemack.livejournal.com]

At street map level, I have the data, but I don't think I've a license to use it for this purpose (conceivably, I have no license to use it for any purpose ...). Also I'm not completely certain of the mapping from latitude and longitude to OS grid references.

At the geological level, this large image (http://www.chiark.greenend.org.uk/~twomack/goo.jpg) is obtained by flood-filling with blue everything connected by a path below 57m to the sea. It's clearly fiction; you don't get islands with that low relief and a sea-bed going down that shallowly, but I don't know where I'd begin in simulating wave erosion. It might be more accurate in the regions beyond the straits of Bath and of Llandogo, where the waves won't penetrate.

The two islands sticking up by the Cheltenham escarpment are the two hills of Gloucester; the limits of the water in the north-east are Kenilworth and Offchurch, in the north-west it flows under Ironbridge and Stoke becomes a harbour town. Chippenham is the little island in the south-east part of the sea; the straits drown Bath. Chipping Sodbury and parts of Filton survive, and there's an isthmus with Portishead on it. I'm afraid Bristol's gone.



The labelled wet image (5M PNG) (http://www.chiark.greenend.org.uk/~twomack/wet-labelled.png)

[tombee.livejournal.com]

Curses, beaten to it.. Bleedin' obvious, really, but extremely cool all the same.

Perhaps of more interest are the strangely regular and non-fractal series of squashed spheres towards the top left of the image. Part of the South Wales Large Hadron Collider, perhaps?

[fivemack.livejournal.com]

They're located west-south-west of Ironbridge, so it's not quite into Wales, and I realise I don't know the name, beyond 'Welsh borders', of that area with Wroxeter, Shrewsbury, Quatford - basically from the Malverns across into Wales. If I had a car, and could figure out a sensible way to navigate along roads towards a GPS position, I'd be very tempted to drive over and see what they look like in ground truth. I suspect they're a signature of a truly spectacular fissure eruption a few hundred megayears ago; it looks as if the spheres were carved out by water escaping from between the dyke and the escarpment.

But I am more ignorant of geography than a worm is of wind-surfing. I've pointed Tim Bond, a geologist at ICL (OK, specialising more in Venus than in Earth at the moment, but one presumably trains long on Earth before proceeding outwards) at this page, and expect him to post something amazingly clear and vastly erudite.