An eighteen-year record of satellite passive microwave SWE estimates from AMSR-E and AMSR2 using the Satellite-based Microwave Snow Algorithm

Abstract:

An eighteen-year record of satellite passive microwave snow depth (SD) and snow water equivalent (SWE) estimates is presented for northern hemisphere seasonal snow. The estimates are derived from the Satellite-based Microwave Snow Algorithm (SMSA), a Japan Aerospace Exploration Agency funded project designed to estimate SD from the Advanced Microwave Scanning Radiometer – 2. The unique aspect of the SMSA approach is that it predicts SWE primarily from satellite brightness temperature observations with the assistance of static landcover data including forest microwave transmissivity estimates (Kelly et al. 2003). In principle, this is a different approach to other methods that use near real-time in situ observations, or SD climatologies to constrain the retrievals from a microwave emission model. SMSA first detects the presence or absence of snow using a standard frequency difference approach before estimating the SD from a constrained dense media radiative transfer model look-up table that requires knowledge of snow density, temperature and effective grain size (Picard et al. 2011). Estimation of these parameters is from simple time-dependent empirical prediction methods (density and effective grain size) and a simple linear model designed to estimate snow temperature. Northern hemisphere SWE and SD estimates are presented for multiple years spanning the Advanced Microwave Scanning Radiometer - EOS (AMSR-E) and AMSR-2 periods of observation. Estimates are tested against in situ SD station measurements, field campaign data and also estimates from the GlobSnow project which have been evaluated with land surface models in a recent study (Mortimer et al. 2020). Uncertainties related to background soil state and snowpack generalised snow microstructure properties are minimized as far as possible with this approach while the method also demonstrates the need for effective tractable forest transmissivity corrections of Tbs. By providing an independent estimation of SD and SWE that is physically-based and with reduced uncertainties compared with empirically-based approaches, the work can help to reduce the knowledge gap regarding global seasonal snow accumulation from space.