Details

Autor(en) / Beteiligte

• Falorni, Giacomo
• Teles, Vanessa
• Vivoni, Enrique R.
• Bras, Rafael L.
• Amaratunga, Kevin S.

Titel

Analysis and characterization of the vertical accuracy of digital elevation models from the Shuttle Radar Topography Mission

Ist Teil von

• Journal of Geophysical Research - Earth Surface, 2005-06, Vol.110 (F2), p.F02005-n/a

Ort / Verlag

American Geophysical Union

Erscheinungsjahr

2005

Quelle

Wiley Online Library

Beschreibungen/Notizen

• The first near‐global high‐resolution digital elevation model (DEM) of the Earth has recently been released following the successful Shuttle Radar Topography Mission (SRTM) of 2000. This data set will have applications in a wide range of fields and will be especially valuable in the Earth sciences. Prior to widespread dissemination and use, it is important to acquire knowledge regarding the accuracy characteristics. In this work a comprehensive analysis of the vertical errors present in the data set and the assessment of their effects on different hydrogeomorphic products is performed. In particular, the work consisted of (1) measuring the vertical accuracy of the data set in two areas with different topographic characteristics; (2) characterizing the error structure by comparing elevation residuals with terrain attributes; (3) assessing a wavelet‐based filter for removing speckle; and (4) assessing the effects of vertical errors on hydrogeomorphic products and on slope stability modeling. The results indicate that in the two sites, relief has a strong effect on the vertical accuracy of the SRTM DEM. In the high‐relief terrain, large errors and data voids are frequent, and their location is strongly influenced by topography, while in the low‐ to medium‐relief site, errors are smaller, although the hilly terrain still produces an effect on the sign of the errors. Speckling generates deviations in the drainage network in one of the investigated areas, but the application of a wavelet filter proved to be an effective tool for removing vertical noise, although further fine tuning is necessary. Vertical errors cause differences in automatically extracted hydrogeomorphic products that range between 4 and 1090.