Temporal Variability of the Satopanth Glacier Facies at Sub-pixel Scale, Garhwal Himalaya, India

Abstract

Growing concern about climate-related glacier change underscores the need to quantify the temporal changes in glacier facies. Classification of glacier facies and assessing their temporal area changes are among the key applications of optical remote sensing in cryosphere. High radiometric resolution (HRR) optical data have an added advantage of overcoming the snow-saturation problem, sometimes observed in medium-to-high-resolution optical sensors. Sub-pixel classification is also known to retrieve the accurate landscape area and addresses the mixed pixel problem in most of the HRR data. Therefore, this paper utilizes the support vector machine (SVM)-based sub-pixel classification approach on bi-temporal HRR data to determine the variability in the surface facies of the Satopanth glacier (SPG), Central Himalaya. Considering the limitations of spectral data in classification, both input Advanced Wide Field Sensor (AWiFS) and reference fine multispectral instrument (MSI) data were aided with the ancillary data like terrain factors, thermal data, band ratios, spectral indices and texture measures. Sub-pixel estimates of SPG facies derived from input AWiFS 2016 image showed good agreement (r >0.7) with their reference MSI-derived estimates. Significant variations were observed in the sub-pixel estimates of SPG facies during the 11-year period (2005–2016). A minimum of ~2% reduction was observed in fresh and slightly metamorphosed snow (FS) area, whereas ice facies showed maximum shrinkage in area (~16%). The maximum expansion of ~8% and ~7% was observed for supraglacial debris (SGD) and ice-mixed debris (IMD), respectively. Wet-snow (WS) and firn coverages slightly increased by ~2 and ~1%, respectively. These changes correspond well with the meteorological data of the SPG obtained from Climate Research Unit Time Series (CRU TS) v.4.01 dataset.KeywordsAWiFSGlacier faciesTemporalSatopanthSub-pixelSentinel