ANNOUNCEMENTS
Accurate mapping of glacier extents can be potentially used to observe glacier area changes, estimate glacier volume changes, understand glacier response to climate changes and monitor natural hazards like GLOF. The study primarily uses InSAR coherence images and combines with elevation and slope data in order to automatically delineate glaciers in Uttarakhand.
The study tests three different methods. The first method takes into account the hit and trial threshold performed on a single coherence image which is used as a primary estimator and slope and elevation data as secondary filters to mask out glaciers. Second method uses multiple coherence images to delineate and extract glacier outlines, trying to reduce the dependency on elevation data for masking out glaciers, whereas the last method uses principal component analysis (PCA) to minimize the redundancy in coherence images and to surpass user intervention (e.g., thresholding, selection etc.) to delineate debris covered glaciers in Uttarakhand.
All three methods are implemented using the Python programming language and their outputs are validated using published outlines of Satopanth and Bhagirath Kharak Glaciers. We find error ranging from 80 meters to 150 meters from these methods. The method output based on PCA shows minimum error.
