Investigation of the landslides with multi-temporal and multi-frequency sar-gnss data: Beyli̇kdüzü-Esenyurt example

The detection and monitoring of landslides which is a kind of natural disaster is very significant. It is required to compare certain conditions with each other in different periods (aerial coverage, velocity of the movement, surface topography, precipation or soil moisture) for determination of present situation and periodic movement of the landslide. The aim of this study is to detect landslides caused by intense activity using InSAR technique and compare the results with GNSS observations in Beylikdüzü-Esenyurt region. In this study, multi frequency satellite sensors (C band, L band) and periodic GNSS observations were employed. In this study, for the entire peninsula between Büyükçekmece and Küçükçekmece lakes 174 Sentinel-1 images with ascending orbit, 169 C band images with descending orbit and 19 ALOS-2 L band images with ascending orbit were used. The evaluations were carried out in the period between January 2015-October 2018 for Sentinel-1 imagery and July 2016- October 2018 for ALOS-2 imagery respectively. Due to the fact that research area is covered by dense settlement area, PSI technique was utilized to determine the landslide behaviors. According to the results, for both orbits of the Sentinel-1, the horizontal movement was determined in the range of -4 mm-3 mm, and the vertical movement was determined in the range of -2 mm-4 mm. The amount of movement determined using ALOS-2 data was higher with the affect of L band. Since the ALOS-2 data was obtained only in one orbit, vertical movement determined in the Line of Sight in the range of -6 mm-10 mm. In order to validate the PSI results obtained in the study, 3 GNSS measurement campaigns were performed in three month periods. When GNSS and InSAR results were compared, it was observed that the movement direction was similar for some points. However, this case was not observed for most of the points. The most likely reason for this case is that GNSS points are not exactly at the same position as the InSAR points and the possiblity of local movements. The landslides show different behaviors in local areas beucause they occur due to affective parameters such as geological formation, precipitation and slope. In accordance with the main purposes of this thesis, the results obtained for the determination and monitoring of landslides reveal the capabilities of multi-temporal and multi-frequency InSAR techniques.