Laboratory-Scale Mapping of Water Percolation in Snow with Upward-Looking Radar and Hyperspectral Imaging

Abstract:

Liquid water percolation in snow is currently not well understood, despite its relevance for snowmelt timing and snow stability. Although it is generally understood that snow microstructure plays a significant role in determining liquid water flow dynamics, there is limited work quantifying snow melt timing coincident with the spatial distribution of water percolation paths for a range of well characterized snow types. Here, we present a series of laboratory-scale snow melt experiments that utilize a novel suite of remote sensing instruments to better understand water percolation in snow. Using a combination of upward-looking polarimetric ultra-wideband (C-band) radar and a compact near-infrared (NIR) hyperspectral imager these experiments aim to improve observational capabilities in the field and the modelling of liquid water flow through snow. From within a controlled laboratory environment, snow samples were prepared with prescribed microstructures, including the presence of capillary barriers and ice lenses, for a range of grain sizes. During snow melt induced with a simulated solar lamp at -5°C, the initiation of melt and timing of the liquid water percolation was tracked non-destructively with upward-looking radar. Following each experiment, change in effective grain size distribution was mapped per 3 mm pixel with NIR reflectance measurements from the hyperspectral imager using serial sectioning (see figure). Additionally, snow microstructure was characterized pre- and post-melt with X-ray computed tomography (micro-CT) measurements. In combination, these techniques allowed for detailed characterization of wet snow metamorphism and water flow features within snow samples, such as preferential flows and horizontal diversions of water due to the presence of capillary barriers and ice lenses.