Snow density retrieval using SAR data: algorithm validation and applications in part of North Western Himalaya

Abstract

The current study has been done using Polarimetric Synthetic Aperture Radar (SAR) data to estimate the dry snow density in Manali sub-basin of Beas River located in state of Himachal Pradesh, India. SAR data from Radarsat-2 (RS2), Environmental Satellite (ENVISAT), Advanced Synthetic Aperture Radar (ASAR) and Advanced Land Observing Satellite (ALOS)-Phased Array type L-band Synthetic Aperture Radar (PALSAR) have been used. The SAR based inversion models were implemented separately for fully polarimetric RS2, PALSAR and dual polarimetric ASAR Alternate polarization System (APS) datasets in Mathematica and MATLAB software and have been used for finding out dry snow dielectric constant and snow density. Masks for forest, built area, layover and shadow were considered in estimating snow parameters. Overall accuracy in terms of <i>R</i>² value and Root Mean Square Error (RMSE) was calculated as 0.85 and 0.03 g cm⁻³ for snow density based on the ground truth data. The retrieved snow density is highly useful for snow avalanche and snowmelt runoff modeling related studies of this region

Separating snow, clean and debris covered ice in the Upper Indus Basin, Hindukush-Karakoram-Himalayas, using Landsat images between 1998 and 2002

“…There could also be some small percentages of steep sloped hanging glaciers, and seasonal snow (re-distributed by wind) at mild slopes but that cannot be separated further. Interestingly, the snow depth at slope 25 o could be up to 10m (Immerzeel et al, 2013), and density up to 100-350 kg/m 3 depending on type of snow/firn (fresh snow has lower density than settled snow or saturated snow or firn) (Cuffey and Paterson, 2010;Thakur et al, 2013). The separate area of perennial and/or seasonal snow zones is thus an important input for glacier mass balance analysis.…”

Snow Density retrieval using Hybrid polarimetric RISAT-1 datasets

“…The climate of the study area is cool and dry. It has three broad seasons, i.e., cold season from October to February with temperature ranging from −15 to 26°C, hot season from March to June with an average temperature between 10 and 26°C, and the rainy season from July to September [18]. It is located at latitude 32.16°N to 32.56°N and longitude 76.91°E to 77.39°E.…”

“…This property of microwave is immensely utilized for snow parameter extraction as most of the snow parameters are correlated with the dielectric properties of the snowpack [16], [19], [20]. In recent years, various studies have been done utilizing SAR-based remote sensing techniques for snow parameter retrieval [15]- [18]. Many physical models based on the theoretical aspects of the wave-target interaction have been used in the past for the extraction of snow properties [15], [16].…”

“…Many physical models based on the theoretical aspects of the wave-target interaction have been used in the past for the extraction of snow properties [15], [16]. The studies using the X-Bragg [17], [18] and the integral equation model (IEM) [19] tried to retrieve the snow parameters such as snow density and snow dielectric constant. The semiempirical modeling approaches were also used for the snow parameters extraction [20].…”

Snow and Cloud Discrimination Using Convolutional Neural Networks

“…With the advent of newly developed remote sensing techniques, a multitude of research works have been carried out to quantitatively assess and monitor the cryospheric elements. These include snow depth and snow water equivalent estimation, measuring snow cover extent, snow wetness, distinguishing cloud and snow and several other related studies (Birajdar et al, 2016;Crane & Anderson, 1984;Deems, Painter, & Finnegan, 2013;Gao, Han, Tsay, & Larsen, 1998;Kulkarni, Singh, Mathur, & Mishra, 2006;Li, Wang, He, & Man, 2017;Man et al, 2014;Mankin et al, 2015;Srinivasulu & Kulkarni, 2004;Tekeli et al, 2016;Thakur, Garg, Aggarwal, Garg, & Shi, 2013;Zhan et al, 2017).…”