The influence of snow microstructure on dual-frequency radar measurements in a tundra environment

• Published in: Remote Sensing of Environment
• Main Authors: King, Josh, Derksen, Chris, Toose, Peter, Langlois, Alexandre, Larse, Chris, Lemmetyinen, Juha, Marsh, Phil, Montpetit, Benoit, Roy, Alexandre, Rutter, Nick, Sturm, Matthew
• Format: Article in Journal/Newspaper
• Language: English
• Published: Elsevier 2018
• Subjects: F800 Physical and Terrestrial Geographical and Environmental Sciences; Tundra
• Online Access: https://nrl.northumbria.ac.uk/id/eprint/34396/; https://doi.org/10.1016/j.rse.2018.05.028; https://nrl.northumbria.ac.uk/id/eprint/34396/1/King%20et%20al%20-%20The%20influence%20of%20snow%20microstructure%20on%20dual-frequency%20radar%20measurements%20in%20a%20tundra%20environment%20AAM.docx

Recent advancement in the understanding of snow-microwave interactions has helped to isolate the considerable potential for radar-based retrieval of snow water equivalent (SWE). There are however, few datasets available to address spatial uncertainties, such as the influence of snow microstructure, at scales relevant to space-borne application. In this study we introduce measurements from SnowSAR, an airborne, dual-frequency (9.6 and 17.2 GHz) synthetic aperture radar (SAR), to evaluate high resolution (10 m) backscatter within a snow-covered tundra basin. Coincident in situ surveys at two sites characterize a generally thin snowpack (50 cm) interspersed with deeper drift features. Structure of the snowpack is found to be predominantly wind slab (65%) with smaller proportions of depth hoar underlain (35%). Objective estimates of snow microstructure (exponential correlation length; l_"ex" ), show the slab layers to be 2.8 times smaller than the basal depth hoar. In situ measurements are used to parametrize the Microwave Emission Model of Layered Snowpacks (MEMLS3&a) and compare against collocated SnowSAR backscatter. The evaluation shows a scaling factor (ϕ) between 1.37 and 1.08, when applied to input of l_"ex" , minimizes MEMLS root mean root mean squared error to less than 1.1 dB. Model sensitivity experiments demonstrate contrasting contributions from wind slab and depth hoar components, where wind rounded microstructures are identified as a strong control on observed backscatter. Weak sensitivity of SnowSAR to spatial variations in SWE is explained by the smaller contributing microstructures of the wind slab.