ASTER: another free source of elevation data

[Holger 0]

Hi there:For anyone familiar with John Childs' excellent DEM website (http://www.terrainmap.com) this isn't exactly news but I haven't found any reference to these free data in this forum and thus wanted to give you a heads-up.ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer, http://asterweb.jpl.nasa.gov/) is a satellite system collecting multispectral images at up to 15 meter resolution between 82N and 82S latitudes. In contrast to the active sensors of the Shuttle Radar mission, the DEMs are derived from "stereoscopic" images of two passes over a given area. According to the ASTER DEM user info (http://asterweb.jpl.nasa.gov/products/DEM.PDF) the resulting accuracies are "up to 7 m absolute horizontal and vertical accuracy". However, the actual accuracy depends on a number of factors, including georeferencing, cloud cover, and the terrain itself, meaning that, at best, the files are similar to the high-res (30-m) Shuttle Radar data. Here's an online article that highlights some of the issues in mountainous terrain: http://www.geo.unizh.ch/~kaeaeb/glims/glim...r-Digital-35882ASTER DEMs are a byproduct of the satellite images and are only generated on request (it seems that anyone can request DEMs to be generated, though the turnaround time is supposedly several months). Therefore, global coverage is not available. However, there are plenty of files on the ftp server and your area of interest might just be among these. Here's a link to the current coverage map: http://edcdaac.usgs.gov/aster/dem_map.htmlFor FS web designers the obvious issues are data quality and data format. As for quality, the main determinant is the cloud cover at the time of image acquisition: clouds interfere with the DEM algorithm and end up being "Swiss Cheese" holes, similar to the holes in the SRTM data, albeit for a different reason. Further, the files cover a relatively small area (60 x 60 km) and even cloudfree areas sometimes have small "ditches" or spikes. Based on my (limited) experience with these data I believe they are best used for local FS terrain meshes or perhaps filling holes in SRTM or other data sets.John Childs' website provides a detailed description of how to order and process ASTER DEMs, including free custom software. The download procedure is pretty awkward but John's step-by-step guide helps immensely. You can execute searches and preview the images and metadata, such as cloud cover and image coordinates. Unfortunately, the skimpy preview images are of the visual bands and it's often hard to see the extent and location of clouds. If you find what you're looking for, write down the file names and download the .tif files directly via WS_FTP or (possibly?) your web browser. For processing I have found that you don't actually have to use John's GEOTIFF4 utility, as both 3DEM and MicroDEM read the ASTER .tif format directly. I use 3DEM (http://www.visualizationsoftware.com/) for previews in 2D and 3D and MicroDEM for the actual processing. For filling the holes, I suggest using DTED0/GTOPO30 data or any other base data you may have.In summary, don't get your hopes up too high, since sketchy coverage and data holes make it difficult to find good files but ASTER DEMs are definitely worth a try, particularly for areas not yet covered by free high-res data. Below is a top-down screenshot of the Swakop River canyon in Namibia, in southern Africa. Note the poor detail of the surrounding area (DTED0 data) but also the sharp edges, even where two ASTER files join. This is due to the fact that most ASTER data are not fully calibrated to the correct base elevation (see the docs on the ASTER website). The other screenshots are taken of the same general area in FS9, with the ASTER data compiled in LOD9 (76-m), i.e. 1 LOD lower than the actual source data resolution. Unfortunately, the default Namibia landclass files are very coarse and often incorrect.Cheers, Holger http://forums.avsim.net/user_files/36371.jpghttp://forums.avsim.net/user_files/36372.jpghttp://forums.avsim.net/user_files/36373.jpg

[Guest vlada stoje]

Hi Holger,thanks for the links, I am looking forward to read also the article about Aster dems. Beautiful pictures! Yes when the web browser supports the ftp, the finished granules are available at ftp://152.61.128.25/pub/asterdem/relative/ or ftp://152.61.128.25/pub/asterdem/absolute/ , I think it depends if there were available the calibrating points for the processing.About the Swiss Cheese holes, look at "my" granule, which I requested 18 months ago (great thanks to John Childs for the explanation!). It seems that there were many clouds during the acquisition of the infrared snaps, but despite this fact I like this data very much. Has anybody some idea about the naming convention of the granules?Greetings from the granule 20020507164545 (???)CheersVladahttp://forums.avsim.net/user_files/36827.jpg

[Holger 0]

Hi Vlada:Did it take 18 months to process your request or is that how long ago you started looking into ASTER data? Have you completed/uploaded any ASTER-based terrain meshes?The naming convention for your '20020507164545' is as follows:2002: Production date (year)05: Production date (month)07: Production date (day)164: Processing strip number545: Sequence number in the processing stripSource: ASTER Users Handbook, page 65 (available at http://asterweb.jpl.nasa.gov/)Even though TERRA (the satellite ASTER rides in) has a similar orbit pattern as Landsat ETM, the naming convention doesn't seem to include any specific row/path identifier like Landsat images have. I guess the online search is the only way to pick out a particular scene. And, yes, watch those cloud cover estimates, particularly if you want to order your own DEM product ;-)Cheers, Holger

[Guest vlada stoje]

Hi Holger,thanks for the explanation of the naming system and for many other precious advices here on this forum! To my granule - it was about two months from the request to the automatic e-mail with the message, that the granule is ready. I am not mass producer, tried prepare just several LOD10 quadrants, with very rough imprecise transformation from UTM to lat/lon (beta version is available at the site VFR Project - Czech Republic visual flying http://vfr.meinlschmidt.org/ , the district "CL" in the northern part of the country). Supposed that the SRTM will be better for this purpose in the future, but it isn

[Holger 0]

Hi Vlada:thank you for your kind words - you're no stranger to giving good advice and sharing interesting news yourself ;-)Beautiful screenshot of the LOD10 mesh! I really like landscapes that are essentially flat save for a few insular mountains (similar to those in the background in the final Namibia screenshot above); it gives a scenery nice definition.As for the geographical corrections, I simply opened the ASTER .tif file in MicroDEM and then exported them as DTED format, with 2 arc-sec (60-m) resolution for my LOD9 file. DTED is a 'unprojected' lat/long format with WGS84 as datum, just what FS wants. Next, I merged the neighbouring scenes with a section of DTED0 data as base (prepared as DTED with 2-arcsec as well). The resulting merged DEM was then ready for processing with bigbsq.exe and the MS SDK utilities.Cheers, Holger

[Guest vlada stoje]

Hi Holger,thanks, it works! The MicroDEM is actually great program, sorry I am not able to understand its full capability. The transformations are pretty fast. It seems that the method Nearest Neighbor is used and it is absolutely perfect for this purpose. The point UTM WGS84 in the zone 33: x=500000 m, y=5600000 m (WGS84 N50.55193 E15.00000) has little bit different coordinates in the UTM on the Krassovsky ellipsoid, which was also used for some maps here (x=500122, y=5600138). Somebody already mentioned also the possibility of this kind of reasons for the shift of different elements in the FS scenery, as the sources can be in different datums and this detail isn

[Guest vlada stoje]

Hi all,another illustration to this issue, the fusion of the Aster elevations with the Srtm30. Sorry no airplane on this screenshot from Surfer, in the FS airplane list we haven

[Guest TMY]

how do you do to convert a file geotiff in DEM?? to be more precise: how do you do to convert in DEM the file geotitt through Microdem?? Thanks Antonio

[Holger 0]

Hi Antonio:quite simple ;-)1. File > Open DEM (select Files of Type GeoTIFF)2. Find your Aster file, then click Open3. The file will load4. You can now subsample the file to make it into a rectangular area of suitable size (remember you have to cover at least one square of the grid of your LOD resolution, e.g. LOD10)5. Also, you can fill any holes with the Edit > Fill DEM holes function or by merging with SRTM30 data.6. If you have Microdem 7, you can save the final DEM as 16-bit BSQ; if you have an earlier version, you will need John Child's bigbsq.exe for the conversion.7. Complete the .inf file, based on the map extent and header information in Microdem, then run resample.exe and the other MS toolsGood luck!Cheers, Holger

[Guest TMY]

OK Holger and tanks bat:>. If you have Microdem 7, you can save the final DEM as 16-bit BSQ; >if you have an earlier version, you will need John Child's >bigbsq.exe for the conversionI have Microdem 7.0 alpha how do I do to save in 16-bit BSQ??? I don't find any voice in the men

[Holger 0]

Howdy:a picture speaks louder...http://forums.avsim.net/user_files/44432.jpgCheers, Holger

[Guest jkvato]

I puzzled over this myself. I had downloaded and installed the complete Microdem 7 alpha version and couldn't find the option to save as a BSQ file. Then I found out that after you install the complete package, you still need to download an updated executable from the following link:http://www.usna.edu/Users/oceano/pguth/mic...32/microdem.zipUnzip that in your Microdem installation folder, and you'll find all the latest options, including the BSQ option.Hope this helps,

[Guest TMY]

OKbut this is version 7.01 and not 7.0 ;-)))))however I have unloaded the new version and all now works Hi and thanksAntonio

[Guest TMY]

30 m - which LOD??? a question: being the data Aster with a resolution of 30 m which LOD needs to use??? the LOD 10 are 38.2 m while the LOD 11 are 19.1 m In the first case the file geotiff results too much small, in the great too second. Is there a solution???? Thanks Anthony V.