Detection of Crevasses in Siachen Glacier Using Remote Sensing Satellite Imageries

Abstract

Snow covered crevasses in the Siachen glacier (Karakoram Range) cause great danger during glacier movement, and the knowledge of their spatial distribution is important for the safe travel. In the present study, crevasses in the Siachen glacier have been detected and further categorized as permanent open/hidden and seasonal open/hidden using a combination of optical (Landsat-8 and Sentinel-2) and microwave (ALOS-2) satellite data. The study is carried out for the year 2018. Initially, the locations of crevasses are manually marked using ALOS-2 data and further, their categorization in open and hidden is done using Sentinel-2 data. Apart from manual marking, the band ratio method is applied on Landsat-8 data to detect the permanent open crevasses in an automatic manner. Gray Level Co-occurrence Matrix (GLCM) technique has also been successfully attempted for the automatic classification of the crevasse and non-crevasse zones. The open crevasse locations using band ratioing are compared with manually marked and a good correlation is found having an accuracy of ~ 93%. A total of 140 crevasse zones have been found within the study area, with 32% as permanent open, 29% as permanent hidden, and 39% as seasonal open/hidden. Manual digitization is done for estimating crevasse dimension (length and width) as this is important input required before the start of any movement. The study reveals that the highly varying nature of crevasses in terms of changes from hidden to open in lower altitude regions of the glacier is mainly observed between July and September. The present methodology of crevasse detection and categorization leads to crevasse information system, which will be used in future to monitor the opening and closing of crevasses in the Siachen glacier on regular basis.