Accuracy assessment of InSAR derived input maps for landslide susceptibility analysis: a case study from the Swiss Alps |
| |
Authors: | Lalan P Singh C J van Westen P K Champati Ray P Pasquali |
| |
Institution: | (1) Geological Survey of India, Hyderabad, India;(2) International Institute for Geo-Information Science and Earth Observation, ITC, Enschede, The Netherlands;(3) Indian Institute of Remote Sensing, Dehradun, India;(4) Sarmap, Cascine de Barico, Purasca, Switzerland |
| |
Abstract: | In recent years SAR interferometry has become a widely used technique for measuring altitude and displacement of the surface
of the earth. Both these capabilities are highly relevant for landslide susceptibility studies. Although there are many problems
that make the use of SAR interferometry less suitable for landslide inventory mapping, it’s use in landslide monitoring and
in the generation of input maps for landslide susceptibility assessment looks very promising. The present work attempts to
evaluate the usefulness and limitations of this technique based on a case study in the Swiss Alps. Input maps were generated
from ERS repeat pass data using SAR interferometry. A land cover map has been generated by image classification of multi-temporal
SAR intensity images. An InSAR DEM was generated and a number of maps were derived from it, such as slope-, aspect, altitude-
and slope form classes. These maps were used to generate landslide and rockfall susceptibility maps, which give fairly well
acceptable results. However, a comparison of the InSAR DEM with the conventional Swisstopo DEM, indicated significant errors
in the absolute height and slope angles derived from InSAR, especially along the ridges and in the valleys. These errors are
caused by low coherence mostly due to layover and shadow effects. Visual comparison of stereo images created from hillshading
maps and corresponding DEMs demonstrate that a considerable amount of topographic details have been lost in the InSAR-derived
DEM. It is concluded that InSAR derived input maps are not ideal for landslide susceptibility assessment, but could be used
if more accurate data is lacking. |
| |
Keywords: | Landslide susceptibility InSAR Digital elevation models |
本文献已被 SpringerLink 等数据库收录! |
|