My notes from reading 'Free and low cost datasets for international mountain cartography'.
My notes from:
Gamache, M. (2004). Free and low cost datasets for international mountain cartography. In 4th ICA Mountain Cartography Workshop (Vol. 26). Barcelona: Monografies techniques, Institut Cartogràfic de Catalunya. [PDF]
- Spikes/wells > 100m were removed
- Oceans set to 0
- Lakes 600m wide set to constant height
- Rivers wider than 183m have been “stepped down in height to highlight them”??
- Tile edge pixels matched
- Filled voids up to 16 pixels
- SRTM has radar angle 31-61 deg, so slopes beyond that angle can be difficult to measure.
- Rabus et al. say there are systematic errors that can be fixed with a single corrective value, though this causes issues with tile merging.
- It’s possible to obtain the raw SRTM measurements
- Patched areas require smoothing, they use focal mean filter with 5x5 window (at 90m)
- Can use satellite imagery using shadow from sun on terrain!
- ASTER is a satellite-based dataset released later than SRTM, they reckon SRTM has better accuracy (which matches my own experience)
- Issues with glaciated mountainous terrain
- Basically useless above 35deg, or aspects between 340 and 140 in northern hemisphere
- Optical DEMs are characterised by a “pit and hammock” pattern (use hillshading for checking)
- Datasets need to be accurately coregistered. Evidence suggests that ASTER and SRTM have substantial shift in places.
- Basically, ASTER has some issues.