[en] Uplift estimation during late Quaternary is required for site selection of geological disposal facility of high level radioactive waste (NUMO, 2004). Terrace level and/or difference in elevation of terraces are good indicators of uplift. Therefore, a reliable method of terrace correlation and chronology is a key issues. Air-photograph interpretation is generally carried out in the early stage of a terrace investigation. However, a terrace classification often depends on the observer's qualitative interpretation. In order to improve objectivity of geomorphic investigation with air-photograph interpretation, we examine to quantify the morphological properties of terrace surface by some morphometric variables that are computed from Digital Elevation Model (DEM). In this study, four morphometric variables (average slope, average laplacian, remaining ratio of a terrace surface, and average depth of erosion) were calculated using data sets of terraces of which chronological data are clearly described. The relationship between these variables and terrace ages shows constant tendencies respond to the geomorphological process caused by the erosion. To examine capability of morphometric variables as an index of terrace correlation, regression analyses were carried out. The regression age estimated from morphometric variables allows to classify terraces in correct sequence, and the error with the observed age falls up to 100,000 years. In addition, to discuss appropriate quantities of DEM for terrace correlation, we used three different elevation data to create DEM: 1) aerial photogrammetry data; 2) airborne laser scanner data; 3) 1:25000-scale contour map. By comparing analysis results of each DEMs, we show suitable qualities of elevation data and DEM grid size to represent the degree of erosion correctly. (author)