Permafrost

Application of ERS InSAR for detecting slope movement in a
periglacial mountain environment (western Valais Alps, Switzerland)
Reynald DELALOYE 1 , Christophe LAMBIEL 2 , Ralph LUGON 3 , Hugo RAETZO 4 , Tazio STROZZI 5
(1.Dept. of Geosciences, Univ. of Fribourg, Switzerland, reynald.delaloye@unifr.ch; 2. Institute of
Geography, Univ. of Lausanne, Switzerland; 3. Kurt Bösch Univ. Institute, Sion, Switzerland; 4. Federal
Office for the Environment, Switzerland; 5. Gamma Remote Sensing, Gümligen, Switzerland)
Abstract: Available data on slope movements in mountain periglacial areas is to date only
restricted to punctual sites. Regional overviews on landslide and permafrost creep (rock glacier)
activity in such areas are often lacking. InSAR (space-borne synthetic aperture radar
interferometry) appears to be a potential tool to fill in, at least partially, this gap. In addition to
the ESA (European Space Agency) SLAM (Service for Landslide Monitoring) program, the
Swiss Federal Office for the Environment (FOEN) intended to test the capability of the ERS
(European Remote Sensing Satellites) InSAR technique for detecting and inventorying both
location and magnitude of slope instabilities in a mountain periglacial environment, namely the
western part of the Swiss Alps. The reliability of the results was evaluated by comparison with
existing and newly acquired field data such as rock glacier inventories, GPS measurements,
air-borne photogrammetry, etc.
ERS-InSAR is sensitive to changes in surface topography in the satellite radar line of sight,
which is inclined at 23° to the east or to the west. InSAR processing generates 2D
“displacement” maps for cm to dm range displacements at cm to sub-cm accuracy. For this
project, 34 interferograms available between 1995 and 2000 have been computed by Gamma
Remote Sensing. Time lapse varies between 1 and 1085 days (only multiples of 35 days +/- 1
day, time necessary for the satellites to be approximately at the same position). Coherent results
can be obtained between early summer and mid fall (the snow free period) and, for daily time
lapse, also in winter when the snow surface still remains cold. They are moreover restricted to
areas located above the tree line or sparsely vegetated. Due to the mountain topography (deep
valleys, steep rock walls, etc.), many gaps occur in image covering.
The major obstacle limiting the potential of InSAR is the presence of wet snow (fresh
and/or old) which strongly disturbs the radar signal. To prevent any misinterpretation of SAR
interferograms, estimating the snow conditions at SAR image dates is necessary in a previous
step. A second significant limitation of the ERS-InSAR technique is that north and south slopes
are not favourably illuminated by the ERS SAR, what decreases the capacity for detecting slope
instabilities in such orientations.
The comparison of InSAR data with “terrestrial” data (air-borne photogrammetry and GPS
survey) showed an encouraging and often excellent fitting of the results in both magnitude
order and spatial pattern of slope motion. The combined use of interferograms with various time
lags (1 day, 1-2 months, 1-3 years) appeared to be a useful step for obtaining a complete
overview of potential slope instabilities in a given area of interest. This systematic method
permitted moreover to identify not only the location and spatial extent of instable zones but also
the magnitude order of the 3D displacement velocities. It was also possible to define typical
ERS-InSAR signature depending on the activity and/or type of different geomorphic landform.
Indeed, more than 600 polygons were identified as ERS-InSAR-dectected slope instabilities
105