SWE Retrieval Analysis Using SnowEx 2017 Airborne SnowSAR Data

Nearly a decade ago, the snow remote sensing community voiced a consensus that no single sensing technique can retrieve global snow water equivalent (SWE) across the wide range of different snow types and confounding factors found on Earth. A multisensor approach is needed. But real world datasets that could be used to quantitatively evaluate how well different techniques worked and under what conditions were rare. NASA's SnowEx campaigns were initiated to fill this gap, and the primary objective of the first SnowEx campaign in 2017 was to collect multisensor observations to enable evaluation of SWE retrieval accuracy in an environment that included the confounding effects of forest cover. The primary sensing technique to be evaluated was radar volume scattering at X and Ku bands, and the European SnowSAR sensor was flown on a P-3 aircraft to collect X and Ku SAR data. Extensive snow ground truth was also collected both in forested areas and in clearings to facilitate the evaluation.

This paper will describe an in-depth analysis of X and Ku-band scattering signatures over the Grand Mesa, Colorado primary site in 2017 as a function of the amount of SWE, as well as the amount of forest cover. We will describe a detailed breakdown of the snow pit and transect ground truth and the corresponding SWE retrieval accuracy using a two-frequency X/Ku algorithm that is at the core of a potential future snow satellite mission concept. We will show detailed examples from the analysis of SnowSAR data, including numerous quality control steps that were necessary at the fine spatial resolution involved. Comparisons of retrieved SWE vs. ground truth SWE as a function of forest cover amount provide answers to the central question from SnowEx 2017, and help lay groundwork for the development of a global snow satellite mission based on the X/Ku radar volume scattering approach.