Semi-automated Technique to Extract Boundary of Valley/mountain Glaciers using Glacio-morphological Information from Digital Elevation Model

Abstract. A semi automated technique has been developed to extract the spatial extension of valleys and mountain glaciers. The method is based on morphological properties of glaciated area extracted from Digital Elevation Model (DEM). Identification of glacial boundary based on spectral information from optical remote sensing imageries produces errors due to misclassification of debris-covered ablation area with surrounding rocky terrain and perennially snow-covered slope with debris free glaciated area. Elevation information DEM of Shuttle Radar Topography Mission (SRTM), CartoDEM and ASTER DEM have been used. A part of western Himalayas was selected as the study area that contains large glaciated basins, e.g., Bhagirathi, Baspa, Chandra basin. First order derivatives, slope aspect, and second order derivatives like, profile and plan curvatures are computed from the DEM. The derivatives are used to quantify and characterise the morphological aspects of the glaciated area and used in the decision rule models to generate the glacial boundaries. The ridge lines of the study areas are also generated from the plan curvature and used in the model to delineate the catchments areas of the glaciers. The slope based boundary is checked for consistency with the boundary from profile curvature and combined manually to generate the final glacier boundary. Area and length under the derived boundary of Gangotri glacier of Bhagirathi catchments are 90.25 sq km and 30.5 km. The result has been checked with high resolution optical data. This objective approach is important to delineate glaciated area, measure the length, width and area and generate glacial hypsometry, concentration factor of the glaciers. Accuracy of the result depends up on the quality of the DEM. DEM generated by SAR interferometric technique is found superior over DEM generated from other interpolation techniques.