Mapping Wet Snow using SAR C-Band through Google Earth Engine

Abstract

Wet snow is an early indicator of snow melting in an area. Increase in liquid water content in snow pack leads to snow wetness. Synthetic aperture radar bands are highly sensitive to snow wetness due to presence of water on snow surface. Spaceborne Sentinel-1 data provides the capability to produce snow maps during spring and summer season which can be used in hydrological and watershed studies. Nagler's wet snow algorithm was applied through GEE in Deosai region which has relatively uniform elevation than surrounding areas and it is covered completely by snow most of the year. Downloading and processing of SAR data is laborious and time taking process which brings the need of a modern solution. Google Earth Engine (GEE) is cloud based, planetary scale remote sensing platform which has a wide range of remote sensing data. GEE has ready-to-use Sentinel-1 GRD product with optical remote sensing data and meteorological data. Algorithm was tested on GEE for study area for the month of June against Sentinel-2 Normalized Difference Snow Index. Wet snow was mapped with an accuracy of 0.8. Methodology was also applied to generate monthly wet snow maps to understand the snow melting period and snowmelt dynamics. Monthly maps of wet snow were generated and wet snow was detectable from March to July in Deosai region.