Author(s): Arnab Bandyopadhyay; P. C. Vanlalnunchhani; Tarak Golom; Aditi Bhadra

Linked Author(s): Arnab Bandyopadhyay, Aditi Bhadra

Keywords: Remote sensing methods; Landsat image; GEI; RGI; Eastern Himalay

Abstract: Accurate estimation of glacier boundary is important as it is a reservoir of freshwater and a primary indicator of climate change. In this study, different remote sensing methods viz. Normalised Difference Snow Index (NDSI), two single band ratios (Red/SWIR and NIR/SWIR) and Automatic Glacier Extraction Index (AGEI) were presented using Landsat 8 OLI image (2017) to acquired accurate information on glacier surface areas in Mago river basin, Arunachal Pradesh. In these remote sensing methods, the first and foremost step was the selection of the best threshold value. Therefore, different glacier boundaries of all feasible threshold ranges (0.4 to 0.9 for NDSI; 2 ± 0.5 for AGEI and single band ratios) for all the different methods were generated. The generated different glacial boundaries were then imported and compared with the Google Earth Image (GEI). In this comparison, inclusion/exclusion of glacier area, shadowed glacier and water feature were checked. It was found that a threshold value of 0.8, 2 and 2, were the best for delineating glacier boundary in the study region by using NDSI, two single band ratios and AGEI, respectively. All the different methods with their best threshold values were imported together and inter-compared each other. This comparison showed that the NDSI captured water feature as glacier whereas the other methods excluded it. Even though both the single band ratios and AGEI were very even, it was observed that the inclusion of glaciers’ area was better in AGEI. Using this method, glaciers in the year 2017 were digitized which was then again cross-checked with Randolph Glacier Inventory (RGI v6.0) and both the digitized glacier and RGI were in accordance. Therefore, from the overall comparison, AGEI with a threshold value of 2 was found out to be the best method for mapping and monitoring glaciers in the study area. Thus, the temporal variations of glacier surface area analysis were carried out for the period between 1988 and 2019. A total shrinkage area of 30.56 sq. km from 1988 (79.97 sq. km) to 2019 (49.41 sq. km) at the shrinkage rate of 1.23% of glacial area per year (0.99 sq. km per year) was observed. During this period the number of glaciers reduced from 55 (1988) to 49 (2019). Upon breaking down the time period as 1988 to 2010 and 2010 to 2019, the retreating rate was found to be 0.78% per year (0.62 sq. km per year) and 2.82% per year (1.87 sq. km per year), respectively. It was observed that the retreating rate was increasing in the recent years. It was also observed that the shrinkage percentage was higher in glacier area smaller than 1 sq. km and lower in glacier area greater than 3 sq. km.

DOI: https://doi.org/10.3850/IAHR-39WC25217119202211

Year: 2022