Appricability of the quantification method II to the discrimination of landslide-prone areas was examined. Six quantities obtainable from a topographical map and likely to be related to landslides, were treated as representative of the geomorphological information. The six quantities are elevation, slope inclination, depth from the summit level, height above the river level, relative relief, and drainage density. Data were collected from each grid square laid on the topographical maps Shinano-Nakajo and Takafu, in which a large number of landslides have occurred. Accuracy of the discrimination was examined by the following four factors. They are precision of elevation measurement on the map, size of the grid square, categorization of the geomorphological items, and data sampling from the grid considered to be safe from landslide risk. The accuracy was largely affected by the data sampling. A simple improvement of the data sampling raised the accuracy of discrimination from 74% to 95%, as judged from a comparison with the actual landslide data.
View full abstract