Determination of Natural Disaster by Integration of Remote Sensing and Gis: the Yeni̇çi̇ftli̇k Stream Basin Model in Istanbul, Turkey

Various methods have been developed to minimise the consequences of natural disasters. GIS and Remote Sensing methods are very attractive, fast and reliable tools for various natural disaster applications and management. In this study, we investigated natural disasters such as flash-floods, floods, erosion and landslide risks, which have occurred, and are likely to do so again, in a study area in Istanbul, to determine the potential use of these methods with respect to these disasters. Disasters which caused loss of life and property in the Yeniçiftlik stream basin, located within the boundaries of Beykoz, a suburb of Istanbul, attracted our attention owing to their particular negative impact on human life and activities. This was selected as the study area. Many geographical parameters, such as vegetation, topographical and geological features, precipitation and land use features played a significant role in the occurrence of other kinds of disasters triggered by flash-floods and flood-related incidents. The data used were topographical, soil, vegetation and geological maps with a scale of 1:25000 (LANDSAT TM (01.08.1987), LANDSAT ETM (28.08.2000), IKONOS (02.03.2008)), and aerial photographs taken in 2006. Geological, soil and vegetation data were converted into raster format by Arcinfo 9.1 software for use in analysis. Land use and vegetation features were determined by the application of a supervised classification technique to IKONOS data. Drainage network morphometric changes were derived from DTM and topographic maps. Land use changes were determined from LANDSAT-TM (1987) and LANDSAT-ETM (2000). Risk maps for landslides, flash-floods, floods and erosion were created in which different weights were assigned to vegetation, geological features, land use, and other geomorphological features, such as slope, aspect, etc.