10:15 AM - 10:30 AM
[HGM03-06] Generation of Terrain Classification Data for Ground Vulnerability Assessment Using 30m DEM by Digital Terrain Analysis
Keywords:Digital Elevation Model, terrain classification, ground vulnerability, landslide, DEM
Digital terrain analysis aims at modeling and estimating various phenomena that are known to be related to topography, using either the terrain measurements (physical quantities such as slope gradient) themselves calculated from a Digital Elevation Model (DEM), or a map that combines these quantities and zones slopes with similar shapes (terrain characteristics). The purpose of this project is to model and estimate various phenomena that are known to be related to topography, and to produce useful data for such modeling and estimation. For example, the risk of slope failure and landslides, the susceptibility of the ground to shaking, the type of soil, the susceptibility to snow avalanches, etc. are known to be related to topography.
In this study, a terrain classification map was created using the Japanese 30-m DEM by calculating the slope gradient, surface texture, and local convexity, which have been found to be effective in zoning medium terrain from previous studies, and the newly added HAND (height above the nearest drainage). The method was simple and reproducible. Using raster images of terrain measurements, the study area was divided into polygons, zoned by clustering with the average value of each terrain measurements in the polygon region as the representative value, and grouped by terrain type in comparison with existing geological maps, landform classification maps, and the distribution of landslides and collapses. The 30-m GeoTIFF file of this classification map and the shape file containing intermediate results such as terrain measurements as attributes are available from the GSI website (https://gisstar.gsi.go.jp/Japan_terrain/).
Through comparison with existing landform classification maps by experts and comparison with areas where shallow landslides and floods have occurred in recent years, this presentation will provide an explanation of the points reached, areas that need improvement, and prospects. In addition, the authors are currently working on the publication of similar data using a global 90-m DEM. This effort will also be presented.
In this study, a terrain classification map was created using the Japanese 30-m DEM by calculating the slope gradient, surface texture, and local convexity, which have been found to be effective in zoning medium terrain from previous studies, and the newly added HAND (height above the nearest drainage). The method was simple and reproducible. Using raster images of terrain measurements, the study area was divided into polygons, zoned by clustering with the average value of each terrain measurements in the polygon region as the representative value, and grouped by terrain type in comparison with existing geological maps, landform classification maps, and the distribution of landslides and collapses. The 30-m GeoTIFF file of this classification map and the shape file containing intermediate results such as terrain measurements as attributes are available from the GSI website (https://gisstar.gsi.go.jp/Japan_terrain/).
Through comparison with existing landform classification maps by experts and comparison with areas where shallow landslides and floods have occurred in recent years, this presentation will provide an explanation of the points reached, areas that need improvement, and prospects. In addition, the authors are currently working on the publication of similar data using a global 90-m DEM. This effort will also be presented.