Automated detection of glacier cover changes in the Upper Indus Basin of northern Pakistan

Abstract

Climate Change impacts start to be more evident all-around, and the cryosphere in the Hindukush, Karakoram, and Himalayan (HKH) region is no exception. Changes in the timing and the amount of snow melt in the Upper Indus Basin (UIB) have water and food security implications for millions of people living in the highlands and lowlands in the region. Previous studies on glacier cover are mostly limited to smaller areas using remote sensing products with a medium resolution, mainly Landsat, with very few of them using automated change detection methods. This creates uncertainty about the evolving state of glaciers in the UIB part of northern Pakistan. In order to address this research gap, we present an automated glacial change detection approach that separates clear ice from seasonal snow and detects changes in clear ice and glacier-covered debris for ~110,000 km2 in UIB of Pakistan. We used a combination of various bands and indices such as Normalized Difference Snow Index (NDSI) and Normalized Difference Vegetation Index (NDVI) retrieved from Sentinel-2 at a spatial resolution of 10m and thermal bands from Landsat 8 to detect and map changes in glacial ice and supraglacial debris. At the same time, limitations of the approach are also underlined, such as threshold selection for such a large, heterogeneous, and complex region that drives future research on the topic. Finally, we juxtapose our results with RGI 6.0 (Randolph Glacier Inventory) and show changes in glacier cover, including clear ice and supraglacial debris, in sub-basins of UIB in Pakistan. Our preliminary results show that the glacier cover has undergone changes in various sub-basins of UIB.

Full text

C44B-04 - Automated detection of glacier cover
changes in the Upper Indus Basin of northern
Pakistan
Abstract
Climate Change impacts start to be more evident all-around, and the cryosphere in the Hindukush, Karakoram, and
Himalayan (HKH) region is no exception. Changes in the timing and the amount of snow melt in the Upper Indus
Basin (UIB) have water and food security implications for millions of people living in the highlands and lowlands in
the region. Previous studies on glacier cover are mostly limited to smaller areas using remote sensing products with
a medium resolution, mainly Landsat, with very few of them using automated change detection methods. This
creates uncertainty about the evolving state of glaciers in the UIB part of northern Pakistan. In order to address this
research gap, we present an automated glacial change detection approach that separates clear ice from seasonal
snow and detects changes in clear ice and glacier-covered debris for ~110,000 km2 in UIB of Pakistan. We used a
combination of various bands and indices such as Normalized Difference Snow Index (NDSI) and Normalized
Difference Vegetation Index (NDVI) retrieved from Sentinel-2 at a spatial resolution of 10m and thermal bands from
Landsat 8 to detect and map changes in glacial ice and supraglacial debris. At the same time, limitations of the
approach are also underlined, such as threshold selection for such a large, heterogeneous, and complex region that
drives future research on the topic. Finally, we juxtapose our results with RGI 6.0 (Randolph Glacier Inventory) and
show changes in glacier cover, including clear ice and supraglacial debris, in sub-basins of UIB in Pakistan. Our
preliminary results show that the glacier cover has undergone changes in various sub-basins of UIB.
Authors
K
Saeed A. Khan
University of Bayreuth
Thursday, 15 December 2022
21:00 - 21:05
Online Only

O
Panagiotis D. Oikonomou
University of Vermont
Colorado State University
Presenting Author
Z
Asim Zia
University of Vermont
University of Vermont
View Related