ASTER GDEM is here!
ASTER GDEM is here!! I blogged about this last year and there has been some eager anticipation over the arrival of the product (see GIM and the BBC). Whilst ASTER has been experimental, I don’t think any remote sensing scientist can same that it has been anything other than a huge success; 14 spectral bands, relatively good spatial resolution and continuous stereo (and so the ability to generate DEMs) make it an almost ideal Earth imaging system. And GDEM is the culmination of 10 years of stereo data collection allowing the creation of a near-global DEM at a 1 arc second (~30m) resolution. A real success.
The data are currently being distributed, freely, via ERSDAC and NASA. The ERSDAC is much easier to use but I never successfully downloaded anything due to (I assume) the load on the server. The NASA WIST interface is (very!) convoluted but works extremely well. You *must* register on the site, ignoring the protestations that the data is offline and that you need to pay. Once through the the full order process you will get a series of emails that will eventually let you download the data via an FTP server. I used the map to define all of Ireland and then grabbed the data which are stored as 1×1 degree tiles; all very easy to do. I then used ERDAS Imagine to mosaic the tiles together.
So what about the data itself?? Well the README is well worth reading, however the main points are land surfaces 83N to 83S, 1×1 degree tiles, 1 arc second resolution, vertical accuracy of 20m and horizontal accuracy of 30m. The file also presents some preliminary QA for this initial release and notes:
“while the elevation postings in the ASTER GDEM are at 1 arc-second, or approximately 30 m, the detail of topographic expression resolvable in the ASTER GDEM appears to be between 100 m and 120 m.”
My initial viewing of the data seems to support this and, generally, I would say that the product is of a lower quality than SRTM (C/X band). That’s in terms of vertical accuracy and resolvable features. 30m C-band or 25m X-band remain generally better products. However there are some things worth remembering: (1) GDEM goes to 83 degrees covering significantly more arctic/antarctic terrain than SRTM; (2) GDEM was collected over 10 years rather than the 11-days of SRTM. Environmental change will therefore be an issue and (3) SRTM is a consistent product that has been refined considerably since its first release. The quality of GDEM will be spatially variable dependent upon the imaging conditions.
It’s great to see this data out which should prove complimentary to SRTM. The coming years should see refinement.