A method to derive satellite PAR albedo time series over first-year sea ice in the Arctic Ocean

• Published in: Elementa: Science of the Anthropocene
• Main Authors: Laliberté, Julien, Rehm, Eric, Hamre, Børge, Goyens, Clémence, Perovich, Donald K., Babin, Marcel
• Format: Article in Journal/Newspaper
• Language: English
• Published: University of California Press 2022
• Subjects: Arctic; Arctic Ocean; Baffin Bay; Canada; albedo; Baffin; Sea ice
• Online Access: https://hdl.handle.net/11250/3057957; https://doi.org/10.1525/elementa.2020.00080

Deriving sea ice albedo from spaceborne platforms is of interest to model the propagation of the photosynthetically available radiation (PAR) through Arctic sea ice. We show here that use of the Moderate Resolution Imaging Spectroradiometer (MODIS) operational surface reflectance satellite product to derive albedo in the PAR spectral range is possible. To retrieve PAR albedo from the remote sensing surface reflectance, we trained a predictive model based on a principal component analysis with in situ and simulated data. The predictive model can be applied to first-year sea ice surfaces such as dry snow, melting snow, bare ice and melt ponds. Based on in situ measurements and the prescribed atmospheric correction uncertainty, the estimated PAR albedo had a mean absolute error of 0.057, a root mean square error of 0.074 and an R2 value of 0.91. As a demonstration, we retrieved PAR albedo on a 9-km2 area over late spring and early summer 2015 and 2016 at a coastal location in Baffin Bay, Canada. On-site measurements of PAR albedo, melt pond fraction and types of precipitation were used to examine the estimated PAR albedo time series. The results show a dynamic and realistic PAR albedo time series, although clouds remained the major obstacle to the method. This easy-to-implement model may be used for the partitioning of PAR in the Arctic Ocean and ultimately to better understand the dynamics of marine primary producers. publishedVersion