Reference : Strategies to estimate ground susceptibility to landslide reactivation. A probabilist...
Scientific congresses and symposiums : Paper published in a book
Physical, chemical, mathematical & earth Sciences : Earth sciences & physical geography
http://hdl.handle.net/2268/5992
Strategies to estimate ground susceptibility to landslide reactivation. A probabilistic application in W Belgium (Oudenaarde)
English
Dewitte, Olivier mailto [Université de Liège - ULg > Département de géographie > Unité de géographie physique et quaternaire (UGPQ) >]
Chung, Chang-Jo [Université de Liège - ULg > Département de géographie > Cartographie et systèmes d'information géographique >]
Cornet, Yves mailto [Université de Liège - ULg > Département de géographie > Unité de géographie physique et quaternaire (UGPQ) >]
Demoulin, Alain mailto [> > > >]
2007
S09-10
Yes
No
International
Society for Mathematical Geology XIth International Congress
3-8 septembre 2006
Université de Liège
Liège
Belgique
[en] Landslide reactivation ; susceptibility ; probability ; fuzzy set membership function
[en] In the hilly region of the Flemish Ardennes in western Belgium, no new big
deep-seated landslides have occurred for decades, whereas several reactivation episodes were
recently observed in ancient landslides. We selected a test area comprised of 13 rotational
landslides located close to the town of Oudenaarde in order to predict the susceptibility of
their main scarp to retreat.
We propose here two probabilistic models based on a fuzzy set approach. The models use
empirical distribution functions (EDFs) as favourability values to build membership values
and combine them by using the fuzzy Gamma operator. Based on Kolmogorov-Smirnov tests
applied to these EDFs to select the most relevant data, a first model was obtained bases on a
combination of 5 quantitative variables: slope angle, distance from cultivation located
upstream of the main scarp, slope aspect, elevation and profile curvature. Another, more
empirical approach based on the a posteriori analysis of the prediction-rate curves was applied
to select the 4 variables of a second model: slope aspect, plan curvature, vegetation index and
focal flow. According to the prediction-rate curves and the resulting susceptibility maps, the
empirical model appears more efficient in locating the main scarp areas most prone to
reactivation.
Researchers
http://hdl.handle.net/2268/5992

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Open access
IAMG06_S09_10.pdfNo commentaryAuthor postprint534.57 kBView/Open

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.