Midday, Thursday Oct 13 2016 local time in New Zealand. A satellite by the name of Landsat 8 passes over the northern part of the Canterbury Plains in the South Island of New Zealand, capturing a near cloud-free image of the area.
The above image shows Kaikoura Peninsula on that October day with its surrounding environs, the ‘Seaward’ Kaikoura Ranges to the North, and the Pacific Ocean to the South-East.
Little did anyone know that one month later, an earthquake would strike, and result in the town being cut off from the rest of the country. People are being airlifted out of the stranded town as the important State Highway 1 route (leading away to the North and South along the coast) is closed due to numerous landslips. The inland State Highway 70 along the valley is also closed, although the officials are trying to reopen it.
Taking a step back, maybe it’s hard to grasp all these ‘geographic’ details I’ve just mentioned, unless you’re familiar with the region. Where is State Highway 1? Why were there so many landslips, is it because there are many steep mountains? How come they’re opening State Highway 70 rather than the coastal State Highway 1?
Part of the problem is the way the information is presented. Quite often, we’re presented with two-dimensional flat maps, or photos that are essentially two-dimensional as well. Worse still, the news might only mention words like the name of the place, a feature like a hill, or that the road-blocking landslip happened ‘somewhere between A and B’.
With a 3D view, you can more easily see why Kaikoura peninsula would be cut off. There’re some freaking steep hills (vertically exaggerated of course) surrounding it! More so to the North (right part of the photo) than to the South (left of the photo).
Also, you can might be able to make out the valley route more easily now (centered in the picture above). That valley is essentially the aforementioned State Highway 70 route where the New Zealand Transport Agency is focusing its efforts on to restore land transport connections to the peninsula.
A picture gives a thousand words because it provides context. What happens when instead, you can interact with that picture?
What if you weren’t just ‘given’ the view. What if you could view it as you please at any angle? Download the data for yourself to play with, make cool new uses out of it,
Give it a try ;)
*More source information available on github.
Some more information (For the more technical minded)
Data-wise, the current 3D demo uses freely available Landsat 8 imagery pansharpened to a spatial resolution of 15m, with an 8m Digital Elevation Model sourced from LINZ Data Service. Special mention to Vincent Sarago @ Remote Pixel for his cool satellite search tools!
Processing was done using an open source Geographic Information Toolstack. Mainly using QGIS, with much help from Luca Congedo’s Semi-Automatic Classification Plugin for satellite pre-processing stages.
Yes, everything here was done completely using free and open source tools :)
What next?
Once the storm clouds clear, hopefully we’ll be able to get a chance to see how the land has changed after the earthquake. If we’re lucky, the new Sentinel 2 satellite might get a shot of the area and we’ll get more photos down to 10m spatial resolution! Granted, that’s not as high a resolution as the Canterbury 0.3m Rural Aerial Photos (2014–15) but it will be much more up to date.
From there on, we’ll be able to do all sorts of change detection. Basically a before and after analysis of the earthquake’s damage. Recent advances in satellite technology is making these sorts of observations all the more easier, with use cases not only limited to disaster response, but also for farmers wanting to know about the health of their crops, scientists measuring the retreat of glaciers and its effects on water availablity, etc.
There’s a lot of potential out there, limited only by your imagination.
P.S. Feel free to contact me for any questions about the data and processing steps.
*Edited magnitude reference from 7.5 to 7.8
*Content originally published on linkedin.
