Automated Retrieval of Snow/Ice Surface Broadband Albedo in Beas River Basin, India Using Landsat-8 Satellite Images and Validation with Wireless Sensor Network Data

Abstract

This paper proposed an algorithm for automated retrieval of snow/ice surface albedo using Landsat-8 satellite images and validation with wireless senor network data in Beas River basin, India. Cloud-free Landsat-8 images of winter season 2016–2017 have been used to estimate snow/ice broadband albedo. Albedo maps generated using satellite images have been compared and validated with in situ recorded albedo values at wireless sensor network stations. Root-mean-square error of 0.053, R-square (R2) of 0.96, bias of − 0.034 and mean absolute percentage error of ~ 3.5% have been observed. Existing global albedo model for Landsat ETM+ was also used to estimate albedo in the study area using Landsat-8 images. Algorithm proposed in this paper for Landsat-8 has shown comparatively better results. Simultaneously, albedo values for snow/ice class have also been estimated using MODIS images. High correlation has been observed between the albedo values estimated from Landsat-8 and MODIS images. Proposed algorithm appears to be the first to estimate broadband albedo of snow/ice class from narrowband reflectances using Landsat-8 images for Indian Himalaya.