Landslide susceptibility mapping in central zab basin using satellite data
Loading...
Date
2014-04
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Teknologi Malaysia
Abstract
Landslides are among the phenomenon natural disasters which cause lots of fatalities and financial losses all over the world every year. The central Zab basin in the southwest mountainsides of West-Azerbaijan province in Iran was chosen as the study area in this research since it is susceptible to landslide due to its climatic conditions, geology and geomorphologic characteristics, and human activities. In order to explain these landslides, various factors, i.e. slope, aspect, elevation, lithology, Normalized Difference Vegetation Index (NDVI), land cover, precipitation, and distances to fault, drainage and road were used. For better precision, higher speed and facility in the analysis, all the descriptive and spatial information were entered into Geographical Information System (GIS) system. In this research, Digital Elevation Model (DEM) was extracted from Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). NDVI and the land cover maps were obtained from the Lansat ETM+ images. Besides, the landslide-inventory map of the study area was identified from SPOT 5. In this study, all the data above input into four statistical methods that include Analytical Hierarchy Process (AHP), Frequency Ratio (FR), Logistic Regression (LR) and Artificial Neural Network (ANN) to produce landslide susceptibility maps. These four methods were validated using R-index analysis and Receiver Operating Characteristic (ROC). Based on the decision support system (DSS), predisposing factors in occurrences of landslides were imported to Automated Land Evaluation System (ALES) software to determine the best options for landslide stability in central Zab basin. The results obtained from the statistical methods showed that, the prediction accuracy were 81%, 86%, 89%, and 92% for AHP, FR, LR and ANN respectively. The map extracted from the ANN model displays higher accuracy as compared to the maps extracted from the other three models. In further study, the findings showed snow melting affected occurrence of landslide in the study area. The importance of snow effect in this basin has also been simulated using a Snowmelt Runoff Model (SRM) as one of the major applications of MODIS-8 satellite images and extraction of Snow Cover Area (SCA) using the Normalized Difference Snow Index (NDSI). The verification results showed satisfactory agreement between the landslide susceptibility maps and the existing data on the landslide locations
Description
Thesis (Ph.D (Remote Sensing))
Keywords
Geoinformation and real estate