Advancing Snow Observation Systems to Improve Operational Streamflow Prediction Capabilities

Objective:

Operationalize low-cost, low-power snow sensing capabilities

Abstract:

In situ snow monitoring infrastructure exhibits limitations for operational hydrologic models in
complex topography because of the spatial heterogeneity of snow distribution. The focus of the
project is prototyping a robust, low-cost, low-power snow monitoring networks for collecting near
real-time, physiographically representative water and energy balance observations in montane
headwater catchments. The sensor networks will complement existing snow monitoring
infrastructure (e.g., SNOTEL), expand edge computing capabilities, and use terrain features and
machine learning (ML) to optimize catchment sensing locations with respect to aspect, slope,
vegetation, and elevation. Site instrumentation will expand on current sensing capabilities to
capture key snow-energy-balance processes, including snow depth, short- and long-wave
radiation, soil moisture, and ML-enabled rain-on-snow (ROS) detection. Using the increase in
catchment observations and new sensing capabilities, the project will benchmark SWE modeling
improvements by comparing physically-based models calibrated using the sensing network
compared to models calibrated with existing snow-sensing infrastructure, using UAV-derived
snow observations and manual snow surveys for validation. Broader impacts of the project will
support an operational pathway to advance snow monitoring networks, provide new datasets to
train and calibrate models, and enhance snow state assimilation into operational hydrological
models (e.g., NextGen) to advance national-scale water resources management capacity.