Eric Sproles | USGS Science

USGS2026

Tracking water in snow: Bridging sensor and and model simulations

Forecasting wet slab avalanches remains one of the most demanding challenges in operational avalanche forecasting. A key obstacle is the difficulty of observing and quantifying liquid water content (LWC) and tracking its infiltration into the snowpack. Field measurements are labor-intensive, highly variable in space, and often influenced by user application or sensor design — making them impractical for forecasting operations. In this study, we evaluated LWC measurements obtained from two capacitive sensors, the SLF SnowPro-17 and -25, and compared them with simulations from the snow cover model SNOWPACK. We analyzed five selected days with pronounced melt events during spring 2025 at Bridger Bowl Ski Area, Montana. The SnowPro-17 measured LWC on the sidewall and the SnowPro-25 was inserted into the pit wall. We then analyzed measurements from the two sensors to assess consistency and potential bias. For each of the five days, we also analyzed liquid water infiltration by layer, comparing field measurements with model-simulated LWC. This allowed us to assess how well the model captured both the timing and magnitude of infiltration processes relative to in-situ observations. Our results show that, the absolute LWC values differ by approximately 1% by vol. on average between measurements from the capacitive sensors, but the timing of liquid water infiltration is largely consistent across both measurements and model simulations. Within our sample, the SnowPro-17 consistently reports higher mean LWC values than the SnowPro-25, and exhibits greater variability, indicated by a larger standard deviation. We found good agreement when we compared in-situ sensor measurements to LWC infiltration in snow cover model simulations. This suggests that, despite measurement challenges and sensor variability, SNOWPACK can provide valuable information on the timing of LWC infiltration. For practitioners, this work highlights that reliable forecasting of wet slab avalanches depends on understanding when and how liquid water reaches critical layers. Because direct LWC measurements are rarely feasible in operational settings, using physically based models may offer a practical pathway forward.

Montana↗

USGS2026

Climate-driven shifts in snowpack conditions favorable for wet snow avalanches: Insights from snow cover model simulations

Forecasting wet snow avalanches remains a major challenge due to their temporal and spatial variability. Long-term climate trends are producing more variable and extreme weather, creating conditions outside historical experience, and increasing forecasting uncertainties. To assess how climate change affects the frequency and timing of wet snow avalanches, we examined the long-term conditions that support wet snow formation. Specifically, we ask: Will the onset of wet snow avalanche conditions or season end shift toward the mid century? And will maximum snow height decrease in avalanche starting zones? To address these questions, we analyze long-term snow cover simulations across our study site in Glacier National Park, Montana, located in an intermountain snow climate. We examined trends in simulated snow depth, and timing of liquid water infiltration over the past four decades (1979–2023) using a physically based snow cover model. To force the model with meteorological data, we use the long-term regional hydroclimate reanalysis dataset U.S. Geological Survey CONUS404 with 4 km spatial resolution and 1h temporal resolution. The use of gridded meteorological inputs to drive a one-dimensional point model introduces inherent scale mismatches. To address this limitation, we use multiple grid points within the study area to simulate a more representative, generalized snowpack, and filter the grid points based on elevation to distinguish between upper- and mid-elevation starting zones. Finally, we analyzed shifts in snow depth, melt-out-day, and LWCindex (bulk liquid water content (LWC) normalized by the residual water content). We use snow depth as a proxy for changes in slab thickness, the LWCindexas an indicator of wet snow avalanche conditions, and melt-out date as a measure of shifts in season length. We ran simulations on flat terrain and on north- and south-facing virtual slopes (38°), across mid-elevation (1,600–1,900 m) and upper-elevation (≥1,900 m) sites for the historical dataset as baseline and a Pseudo-Global-Warming (PGW) dataset to identify a warming driven shift. Across all aspects and both elevation bands, PGW forcing produces a consistent reduction in annual maximum snow depth (median shift −0.15 to −0.26 m, p<0.001) over 44 seasons, with the largest reductions on north-facing slopes and a modest but significant strengthening of this reduction over the 44-year simulated record. Melt-out starts 10–14 days earlier under PGW across all aspect–elevation combinations (p<0.001), again most pronounced on north-facing, higher-elevation terrain. The LWCindex threshold-crossing dates likewise shift earlier under PGW, indicating an earlier onset of wet snow avalanche conditions across the domain. Together, these results point to a snowpack that is both thinner and wetter earlier in the season under future warming, with the strongest signal consistently on north-facing slopes at higher elevation. This work lays the groundwork for climate-informed forecasting and mitigation and offers historical context for avalanche operations in a changing climate. Understanding long-term trends can help avalanche professionals interpret current observations, improve wet snow avalanche forecasting, and develop more effective mitigation strategies in a warming climate.

Montana↗