Investigating the potential of snow liquid water content retrieval from near-infrared reflectance measurements

Assessing snow melt and the liquid water content (LWC) of snow is crucial for understanding the hydrological cycle for predicting water resources, hydroelectric power generation, runoff, and potential flooding. It is also essential for correcting remote sensing signals (RADAR) and forecasting wet snow avalanches, for which snow stability is closely linked to its water content. Various methods exist to measure snow LWC, including calorimetry techniques, centrifugal separation, and dielectric methods based on permittivity differences between ice, air, and water. While these methods are well established, they are limited to low sampling resolutions and do not capture the typically high spatial variability of liquid water within the snowpack. However, Donahue et al. recently (2022) demonstrated the potential of near-infrared (NIR) spectral imaging for visualizing the 2D spatial variability of snow wetness in their study, Mapping Liquid Water Content in Snow at the Millimeter Scale: An Intercomparison of Mixed-Phase Optical Property Models Using Hyperspectral Imaging and In Situ Measurements (The Cryosphere).

The SnowImager instrument (snowimager.ch), recently developed at the Institute for Snow and Avalanche Research (WSL/SLF) together with a local start-up (Davos Instruments), allows for measuring the 2D spatial NIR diffuse-reflectance of snow stratigraphies at wavelengths of 850 nm and 940 nm. Leveraging the fact that reflectance at 850 nm is less influenced by liquid water than at 940 nm, we explore the application of NIR diffuse-reflectance imaging for measuring 2D LWC distribution with the SnowImager. As a first step, we developed a wetness index based on the reflectance measurements, and which is proportional to the LWC. Because the NIR diffuse-reflectance also depends on the optical equivalent grain diameter, a baseline dry reflectance ratio was determined using dry snow samples collected over the winter season 2023/2024. In addition, field measurements (in Weissfluhjoch test site and in Tschuggen during the melt season) were carried out to compare the wetness index against conventional liquid water content measurements obtained with a capacitive sensor.

Results from the Tschuggen campaign exhibit good agreement between the wetness index and the LWC measurements with the capacitive sensor for the snowpack wetness evolution. Furthermore, the imaging approach demonstrates the ability of capturing high resolution 2D variability of the LWC within a snowpack. Although the findings are promising, limitations were identified at snow microstructure regions of high textural contrasts. Further research is required to validate the wetness index method comprehensively, particularly concerning the characterization of the baseline reflectance ratio.