Trekking the High Himalaya for Data
This post covers how a range of different remotely sense data sets can be acquired from around the world with a focus on Nepal
In a recent post, we looked at how Open Street Map can be used to acquire some basic geospatial data for many parts of the world. When we left off, we were still looking for some elevation data for Krishna and came upon the possibility of downloading contours from OSM as seen below on the Cycling map:
With a bit more digging, we determined that this wasn’t possible – but we were able to figure out where the contour data came from. And better still, we figured out how we could get a copy of it. From the OSM Wiki, we determined that the contours were derived from a DEM that had been derived from data collected on a space shuttle Endeavour mission in 2000. The mission aim was to map the majority of the earth’s surface (~80%) using radar, thus enabling the creation of an almost complete global DEM. This was the Shuttle Radar Topography Mission (SRTM). In the shuttle’s payload bay was a radar transmitter, sending out a strong radar signal. Also in the bay was a receiver. A 60 metre boom extended out from the payload bay with another receiver at the far end of it. The radar reflections were received by the two antennae; the separation between the two allowed for a stereoscopic view of the earth’s surface which, after some pretty grunty computations, resulted in a reasonably high resolution DEM.
And here’s a map of the coverage. The colours indicate the number of times a particular area was covered (red means no coverage).
Not too shabby. This being a US funded project, SRTM data for regions outside the United States were sampled for public release at 3 arc-seconds, which is 1/1200th of a degree of latitude and longitude, or about 90 meters. Those areas in the US were released at 1 arc-second, or roughly 30 m. Last year, NASA agreed to release the 30 m data for the rest of the world so things have suddenly gotten a whole lot finer. A selfish look at the figure above indicated that Nepal had been covered at least once, so there was hope for Krishna. Some more digging got us to the US Geological Survey’s Earth Explorer website:
From here, we could pan the map over to the area we were interested in and then search for which data are available. This is a warehouse of remotely sensed data so we’ll have a very quick look at what’s available, with a particular interest in elevation data. Having navigated over to the Mustang district, we could use the cursor to click the four corners of a box around our area of interest:
Next, we’ll click on the Data Sets tab and see what’s out there.
(One got cut off at the bottom – it’s vegetation monitoring.) We could spend a few days going through each of these (and I invite you to browse) but for now let’s focus on the Digital Elevation entry. When expanded, it looks like this:
So we can clearly see some SRTM entries, as well as a number of others. If we stay focused on SRTM for now, we’ve got four to choose from: 1 Arc-Second Global (at ~30 m), Non-Void Filled (90 m), Void Filled (90 m) and Water Body Data (90 m). Clicking on the “i” takes you to a page with more information (metadata) on that layer. I’m pretty keen on that 1 Arc-Second data so I’ll tick the box next to it and hit “Results” at the bottom of the page. We got a nice hit off of that – clicking the footprint button, , displays its extent on the map (the larger, light red box) – we’re covered! Next step is to download this. You’ll see the “download options” button to the right of the footprint. Note that you’ll need to register to get a username and password, but that’s no big deal.
Here we’ve got three options. Both BIL and DTED will require a bit of pre-processing before we can use them, but a GeoTIFF we can use almost straightaway so let’s grab that.
A GeoTIFF is an image format that also has geographical data built in, particularly a coordinate system. This means that when I add it to a map, it will automatically be placed in its correct position (depending on the system), which is quite handy. And even though it’s a TIFF, we can work with it as if it were a regular raster grid that we’re more used to working with. Here it is added to an ArcMap map:
Note the coordinates in decimal degrees (at lower right) and elevations ranging from 850 to 5033 m (in the legend). Here’s a hillshade derived from this:
Almost there – before we can do much with this grid (e.g. slope, aspect) we need to project it from its current system, which is a geographic coordinate system (WGS84 to be exact, the one used by GPS) to a projected one. So here goes.
- In ArcToolbox, go to Data Management Tools > Projections and Transformations > Raster > Project Raster;
- Set your GeoTIFF to the input raster – note that the coordinate system is picked up automatically;
- Note where the output is being saved – change its location and name if necessary;
- For the Output Coordinate System, click the button;
- Navigate to this file: Projected Coordinate Systems > National Grids > Asia > Nepal Nagarkot TM (This is a local projected system for Nepal based around a place called Nagarkot, east of Kathmandu):
- Click OK and the new grid will be added to your map.
With this projected grid we can do any sort of raster analysis we like now. And it didn’t cost the world to get it. You’ll note those bright areas – which my best guess would be snow. Radar reflectance off of snow (and liquid water) is pretty minimal and provide no reflection, so the quality of this DEM may have been affected by when during the year it was acquired. With this in mind, I went back and looked at the other data sets. The ASTER data (Advanced Spaceborne Thermal Emission and Reflection Radiometer) (Egads!) is another remotely sensed DEM at a 30 m resolution which boasts 99% coverage of the earth’s landmass. Downloading that one as a GeoTIFF and projecting it as above got me this (I’ve called it Jomsom_ProjectRaster.tif after a town near our area of interest):
This is a much more complete DEM – we now have elevations up to 8,147 m. And here’s the hillshade:
Let’s add the vector data we got from Open Street Map previously and see how it all fits together.
The hillshade’s not quite rockin’ to my satisfaction but the more important thing at this point is that we’ve now got some data to work with, and all it’s really taken is time. We’re still lacking in crucial data such as land cover or population, let alone climate or soils, but it’s a start, and a lot better of a start than I imaged when we started looking into this.
I often suggest to students that if GIS is an engine, then data are the fuel. Any spatial analysis will be limited by the availability (and quality!) of the data used, and in many cases, the data limit what can and can’t be accomplished. For many parts of the world, particularly in developing countries, it is a real challenge to get any usable data. In Krishna’s case, some of the crucial data he needed was found off the internet and I think we did reasonably well in that respect. In future posts we’ll explore other sources of global data, but I hope this post and the previous one have given you a sense of where to start looking when you’ve got no data at all to start with.