When I actually work with SRTM 3 data in a meteorological application and especially when doing simulation over mountainous regions I expected and found a lot of voids in SRTM3 data. Has there been anyone here who has actually fixed the voids by interpolating using bilinear interpolation as this developer has done - http://fossies.org/dox/gpsprune_17.1_source/LookupSrtmFunction_8java_source.html and then can talk about their experiences with the "fixed" SRTM 3 data ?

Or the better option is to download the NASA void filled SRTM 3 data. I do not want to go down the route of using ASTER GDEMs for various reasons.

  • Isn't the easiest option to download NASA's SRTM v3 data? Why would you want to go down the difficult path of implementing void-filling yourself when NASA has already done the job (and probably better than something you could do)? – Igor Brejc Nov 10 '14 at 13:03
  • @Igor - I agree downloading would be best. But I was just asking for user experiences in this field especially in regards to mountainous regions. – gansub Nov 10 '14 at 15:59

There is a SRTM v4 version with gaps filled using the method described in Reuter et al (2007). It is availalbe on the cgiar Website. Note that users are prohibited from any commercial, non-free resale, or redistribution without explicit written permission from CIAT.

Personnally I've found that the version from Jonathan de Ferranti was slightly better (only tested in Europe) because it is filled and corrected with ancillary data. But there is no published method.

Reuter H.I, A. Nelson, A. Jarvis, 2007, An evaluation of void filling interpolation methods for SRTM data, International Journal of Geographic Information Science, 21:9, 983-1008.

  • 1
    SRTM v4 (from cgiar) = licence due to processing can affect you (so depending on your use case, not a good idea) srtm.csi.cgiar.org/SELECTION/SRT_disclaimer.htm. See also wiki.openstreetmap.org/wiki/SRTM – ThomasG77 Nov 10 '14 at 13:34
  • I've mentioned CGIAR because their method is documented and published, but as I said and used the seond dataset (not because of the licence because it was for reserch, but because the mountain areas were better). Antway, you are right that I should heve mentioned the licence, so I've updated my post. – radouxju Nov 10 '14 at 13:56

Using the data on http://e4ftl01.cr.usgs.gov/SRTM/SRTMGL3.003/2000.02.11/ seems like a good option. Does anyone know if this site will be maintained?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.