Application of Satellite Images to Studying Correlation between Landslides and Land Use in the Taipei Metropolitan Area
Date Issued
2010
Date
2010
Author(s)
Lin, Chih-Yung
Abstract
This research deals with the correlation between landslide hazards and land use types in the Taipei metropolitan area, northern Taiwan. Research approach is mainly use of Formosa Satellite images and land-use survey data in 2007. Satellite images were collected in the winter of 2006, the summer of 2007 and the winter of 2007. The NDVI were extracted from the images collected in the 2006 and 2007 winters. In order to secure accuracy, misinterpreted landslide samples were deleted by comparison with supervision classification images, cloud shapes and slope data.
With annual variation data available, this research proceeded in two directions. One direction is environmental characteristics of landslides in the Taipei metropolitan area: slope, elevation, aspect; the other is the correlation between land-use types and landslides. Normalized statistical data show that most landslides are in eastern, southeastern and southern aspects as well as in slopes I and slope II (0%-15% slope) with elevations below 500 meters. Five categories of land use, i.e., forestry, transportation, built areas, mining and natural land, are in close correlation with landslides, while agriculture is not and this may be due to significant variation in physical attributes and wide distribution of farming land.
In addition, analysis for smaller areas is based on government-defined four clusters in terms of landslide risks. The research finds that built areas and transportation facilities are obviously correlated with landslides in potential landslide zones with mid-high, mid-low and low risks. The government should be careful for land development in such landslide-correlated areas .
Subjects
Correlation test
Formosa Satellite Images
Hazard
Land Conservation
File(s)![Thumbnail Image]()
Loading...
Name
ntu-99-R97228009-1.pdf
Size
23.53 KB
Format
Adobe PDF
Checksum
(MD5):7a1ae9fcf878e0a47b57deeaa91eed9c