REMOTE SENSING STUDIES OF CHHOTA-SHIGRI GLACIER, HIMALAYAS

Abstract

The Himalayas, designated as the third pole of our Earth, is the land of thousands of glaciers. These glacial snowfields store about 12,000 km3 of freshwater, which is the supply source for perennial rivers such as the Indus, Ganga and Brahmaputra feeding most part of India. Studies have shown that glaciers are shrinking due to the effect of climatic change. The changes in the glaciers may lead to catastrophic events, viz. glacial lake outburst floods (GLOF). These changes need to be monitored periodically to assess the glacier health and prevent such events. Contemporaneously, it is observed that the knowledge about the glacial changes is very limited over the Himalayas. In India, the glaciological studies have been limited to primarily field based studies but it has been emphasized in several studies that the synoptic nature of remote sensing can play a major role in completing the inventorying of the glacier database. In today’s world, large temporal database of remote sensing images are available, which can be used to monitor the past changes also. In this research, using field data, satellite images and ancillary data (toposheets, previous glacier maps and Digital elevation models (DEMs)), several aspects of Chhota- Shigri glacier have been studied. Field data includes collection of GCPs, temperature information and photographs of different glacial features. Satellite images of varying spatial, spectral and temporal resolution from ASTER and WorldView-2 PAN sensors have been used to generate credible data about the glacial landforms, terrain cover, surface ice dynamics and recession of this glacier. DEMs of varying spatial resolutions (ASTER DEM and SRTM DEM) have been used to extract primary and secondary topographic parameters. Data management, processing and analysis have been done in the GIS platform. Digital image processing techniques such as orthorectification, supervised classification, band transformation, edge enhancement and image correlation have been extensively used in this research for glacial terrain mapping and dynamics study. ii A novel methodology has been proposed for the mapping of debris-covered glacier terrain classes and the results have been compared with the output of maximum likelihood classification (MLC). In the developed methodology, a new hybrid classification method has been proposed which integrates both MLC and knowledge based classification (KNC). Empirical formulae based on band transformations for contrast enhancement of ASTER bands have been proposed, which have increased the separability of various glacier cover classes, especially between cloud and snow. The classified image from the MLC and the transformed bands has been found to be very accurate with an overall accuracy of 89.75%. The user's and producer's accuracies of supraglacial debris have been found to be 91.40% and 92.10% respectively and for periglacial debris 91.25% and 92.23% respectively. Cloud has also been classified accurately using the transformed bands which justify the contrast enhancement performed to increase the separability between the glacial terrain classes. An evaluation of MLC vis-à-vis hybrid classification has been done. Hybrid classification gives an overall accuracy of 86.65% which is above the acceptable values. The user's and producer's accuracies of supraglacial debris have been found to be 85.50% and 86.50% respectively and for periglacial debris 89.63% and 90.10% respectively. Though the individual accuracies for supraglacial debris and periglacial debris have decreased in comparison to MLC based classification, yet they are above the general acceptable values. This classification does not require post-processing of the output maps since the user knowledge about the classes are inherently taken care of during the classification. The proposed method however needs to be tested in other larger areas or with different debris-covered glaciers with diverse debris properties for establishing its utility in glacial mapping. The effects of changing climate are well seen as markers on the glacier surface as well as in the valley walls which embrace them. Field based methods give insight to the change in the glacier surface but they are limited due to highly precarious terrain. The synoptic nature of remote sensing in conjunction with field data can help overcome these limitations. Considering this, a large scale glacial geomorphological map of the glacier has been prepared using satellite images and ancillary information. The DEM and DTM of the area along with the high spatial resolution WorldView-2 multispectral and panchromatic images have been found very elucidating particularly for the landform distribution. iii The recessional trend has been studied for the Chhota-Shigri glacier. Past and recent recessions have been estimated using different satellite images, time series ASTER images and ancillary data (DEM and DEM derivatives). Paleo-morainic loops have been mapped in the inactive zone of the glacier, presence of which is the indicator of recessional history of the glacier. The textural differences seen on the high resolution WorldView-2 multispectral and panchromatic images in conjunction with the surface elevation profile and the DEM derivative like rate of change of slope have been used to map the exact position of these loops. Seven distinct loops have been mapped using this technique and their locations have been verified using the published map of loops, which were prepared using field surveys. A good correlation has been found between the two. The map prepared using field survey shows presence of only five distinct loops whereas using the proposed approach seven loops have been mapped. This clearly indicates the advantage of using remote sensing technique over the field based surveys in perilous terrain like Himalayas. Surface ice velocity of this glacier has been derived using sub-pixel image correlation technique (COSI-Corr software) on the ASTER time series images (2003– 2009). The remote sensing derived measurements have been found to match quite well with the field measurements. In general, the surface ice velocity varies from ~20 my-1 to ~40 my-1. Velocity variations occur in different parts of the glacier and also from year to year. In time period considered for this glacier, the mid-ablation zone and the accumulation zone exhibit higher velocities whereas the zones near the snout and equilibrium line altitude have relatively lower velocities. Further, the velocities have been found to be relatively higher in the years 2005–2006 and 2007–2008 and lower in the years 2006–2007 and 2008–2009. These spatial and temporal variations in velocity, which may be related to the glacier morphology and hydro-metrological factors, need to be further studied. The studies conducted in this research fairly demonstrate the synergy between field based and remote sensing based methods for mapping different aspects of glacier. All the results derived in this study have been validated using field data or the published maps based on field surveys. The high correlation between the field data and the remote sensing based derived results has been observed in this research. The large available database of the satellite image makes it an effective tool for filling the knowledge gap in the field of glaciology. The research reported in this thesis corroborates the use of remote sensing iv techniques to map the glacial extent, changes in the landforms, glacial recession and for deriving long term surface ice velocity data with high accuracy.