Supervisors info:
Μαρουκιάν Χαμπίκ Ομότ. Καθηγητής (Επιβλέπων), Παπαδοπούλου Κυριακή Καθηγήτρια, Γκουρνέλος Θεόδωρος Καθηγητής
Summary:
In the past the area of Pilio (county of Magnesia), especially the urban areas
and the road network, has been repeatedly affected during the past, by
landslides. The purpose of this study is to produce a landslide susceptibility
map for the study area by using Geographic Information Systems (GIS). The
landslide locations were identified using past references and field surveys.
There are 8 landslide inducing parameters considered for the completion of the
landslide susceptibility map. These parameters are geology and distance from
lineaments, slope and aspect, distance from streams and roads, land uses and
precipitation amount. The selection of these factors was based on the observed
correlation with landslide occurrence in the study area. Landslide
susceptibility was then analyzed, using all these landslide occurrence factors,
by employing the Frequency Ratio model. The Landslide Susceptibility Index was
graded into four classes using the Quantile method. From the created landslide
susceptibility map can be observed that the highest rates of Landslide
Susceptibility are observed in the northeastern part of the area where the
communities of Tsangarada, Kissos, Zagora and Pouri are settled. In this area
the lithological formation of schist is prevailing, the precipitation amount is
higher and the slopes are steep. Whereas the lowest rates of Landslide
Susceptibility are observed in the central and southern Pilio where the
communities of Neochori, Afetais and Milina are the most common. In this area
the lithological formation of marbles is present, the precipitation amount is
lower and the slopes are more gentle. For the verification of the analysis
results and the estimation of the accuracy of the landslides prediction, random
landslides have been used that they had been excluded from the landslide
susceptibility analysis. These random landslides along with the results of the
landslide susceptibility analysis have been used in order to perform a ROC
(Reciever Operating Characteristics) curve analysis and to estimate the AUC
(Area Under Curve). The observed accuracy is equal to 78,2%. The produced map
can be very useful by identifying roads and areas prone to future landslides.
In this way, correct decisions for future land use and major construction
projects can be made.
Keywords:
Landslide, Susceptibility, Pilio, Magnesia, GIS
