Mapping Snow Sensor Usability in the Northern Hemisphere with Google Earth Engine

Abstract

Remote sensing provides a powerful tool for regularly observing seasonal snow properties across local, regional, and global spatial scales. Satellite Passive Microwave (PM) remote sensing provides a record of over 40 years of observation of snow properties like snow depth (SD) and snow water equivalent (SWE). PM sensor retrieval of snow can, however, have errors and uncertainty due to vegetation cover, snow depth, and snow wetness. While these limitations have been well-studied, they have not been organized to inform the application of snow products for other fields of research and/or resource management. This paper presents “Snow Sensor Usability Masks” (SSUM) that provide classifications where PM has demonstrated capability, potential capability, or no capability based on results from peer-reviewed publications. During the snow season (October to April), 33% of snow-covered areas in the Northern Hemisphere (excluding Greenland) have demonstrated capability with PM sensors. January has the greatest capability (42%) in the Northern Hemisphere, with February following closely (37%). As a case study, evaluation near Quebec, Canada for the month of February illustrates that capability increased more when forest canopy thresholds increased than when SWE thresholds increased by order of magnitude of two. Our findings support the need for further development in methods to detect and quantify snow beneath forest and vegetation in PM radiance assimilation. This paper provides guidelines for applying PM snow products across the globe, as well as a framework for setting priorities for future PM data assimilation algorithm development and future snow field campaigns.