Assessment of Mass Movements Driven by Exceptional Meteorological Events Throughout the Use of DInSAR in Austria
Vecchiotti, Filippo; Kociu, Arben; Tilch, Nils; Winkler, Edmund
Austrian Geological Survey (GBA), AUSTRIA

Introduction

The main focus of this paper itīs to elaborate a cost effective strategy for ERS and ENVISAT images processing in the aim of assess the Austrian mass movement cadastre through the use of open source software.

Methods

The methods used for a cost efficiently use of the archived SAR images was structured in the following way:

  • A predictability PSI layer for ascending and descending orbits by using CORINE data and DEM derived slope and aspect map was created and compared to the GEORIOS landslide cadastre of Austria.

  • A series of wheatear information and snow data from the HYSTALP project allows for the discrimination of the probability to find snow free areas in specific months of the year
    (July and August for the high mountains domains, August and September for hilliest areas, whereas March, October and November for valley resorts).

  • By reviewing historical rain, snow, flood and landslide events happened on the time frame 1992 - 2010 it was possible to apply a so called ''pre-triggering, co-triggering and post-triggering'' method
    in order to verify if with a two pass DINSAR technique the major meteorological event activates new or re-activates previous mass movements in Austria.

  • Furthermore local weather rain stations (over 1200) allows for the right choice in terms of date of image acquisition; being the humidity a factor which introduces a high atmospheric noise on the radar signal.

    The proposed method aims at verify for the whole scenes the best compromise between statistical probability to detect movements on the ground, temporal and geometric decorrelation,
    attenuation of the electromagnetic wave by the atmosphere for DINSAR products.

    Analysis

    New functionalities of NEST open source software by processing two couple of images per site in three different time periods itīs been tested in 3 sites (high mountain relief, medium mountain relief and valleys).
    The final interferograms were furthermore unwrapped using the SNAPHU algorithm in order to obtain a series of map expressing the amount of displacement in millimeters.
    The comparison of the DINSAR displacement maps with the GEORIOS cadastre, geological, geomorphological and geotechnical available data were performed in order to investigate precursory signals connected to the activation of major rapid catastrophic mass movements.

    Results and conclusions

    The ability to detect very slow deformation phenomena like deep seated landslide, creep and superficial loose soil landslides together with subsidence induced by groundwater level fluctuation and collapsible ground was positively assessed.
    However the possibility to associate precursory vertical displacements to on-going slope degradation processes leading to catastrophic mass movements
    requires a more detailed advanced DINSAR investigation expecially if the real objective is to deeply understand the mechanisms behind the detected phenomena.