The Version 1 AMSR-E/2 Rain-on-Snow Data Record (ROS) was generated from daily brightness temperatures (Tb) at vertically and horizontally polarized 36.5 GHz and 18.7 GHz from the Advanced Microwave Scanning Radiometer for EOS and Advanced Microwave Scanning Radiometer 2 (AMSR-E/2) across Alaska, USA for water years (WY) 2003 to 2016. The AMSR-E/2 orbital swath Tb data were spatially re-sampled to a 6-km resolution on a polar EASE-Grid (version 2) format using an inverse distance squared weighting method (Brodzick et al. 2012, Du et al. 2017). To ensure cross-sensor consistency, the gridded AMSR-2 Tb data were empirically calibrated against the same AMSRE frequencies using a Double-Differencing method and similar overlapping observations from the FY3B MWRI sensor record (Du et al. 2014).

ROS events and associated snow wetness, induced by atmospheric conditions, were detected using a spectral Gradient Ratio to exploit the 19 and 37 GHz dielectric properties in response to rain-on-snow events and enhanced liquid water content (LWC) within the surface snowpack (Grenfell and Putkonen 2008). The Gradient Ratio is applied to vertical and horizontal polarizations, respectively. The proceedings Gradient Ratios are then ratioed together using a Gradient Ratio Polarization (GRP) (Dolant et al. 2016, Langlois et al. 2017). A threshold value is then applied to the GRP to detect ROS. The threshold is set to < 1 for elevations below 900 m and < -5 for elevations above 900 m.

The satellite derived data record allows for the identification of ROS events at a relatively fine spatial resolution and high temporal resolution. Data are provided at daily acquisitions, allowing for the identification of single ROS events; and monthly annual aggregations from November through March.