Linking Variations in Gangotri Glacier Features Mapped at a Large-Scale with Climate Variables

Abstract

Changes in glacier area, glacial lakes, debris cover, and geomorphological features such as debris fans have a significant impact on glacial dynamics. Therefore, precise and timely observation and tracking of glacier surface changes is a necessity. The availability of high spatial resolution remote sensing images has made it viable to analyse the glacier surface changes at a local level. However, with increasing spatial resolution, the spectral variability increases, giving rise to additional challenges (such as false changes and misregistration) in the change detection process. These challenges can preferably be dealt with using object-based change detection approach than the conventional pixel-based change detection approach. In this study, a class-object change detection approach has been followed in which the independently classified objects from multi-temporal images are compared to detect the changes. As a case study, the changes in Gangotri Glacier (Uttarakhand Himalayas) features have been studied using high spatial resolution WorldView-2 image and Linear Imaging Self-Scanning System (LISS)-4 image for a three-year period 2011–2014. The spectral correspondences between glacier surface and non-glacier surface have been handled by considering brightness temperature and slope as ancillary data to improvise their distinction. A change detection accuracy of ~ 84% has been obtained using the OBCD approach.