Weather Radar Retrievals of SWE

Snow water equivalent (SWE) estimation is crucial for water resource management in snow-dominated regions, but current measurement technologies have limitations. This study introduces a novel approach combining dual-polarization weather radar technology with machine learning algorithms to enhance SWE estimation and improve hydrological forecasting.

Our research leverages recent advancements in weather radar detection algorithms, specifically dual polarization radar, which improves the characterization of ice water content (IWC) as particles descend through the atmosphere. We address the challenges of translating IWC flux aloft to ground-level SWE by employing SnowQ™, a patent-pending method utilizing machine learning.

We analyzed three years of weather radar data from the Conejos River Basin in Colorado's Southern Rocky Mountains. Our machine learning techniques were trained on high-resolution airborne LiDAR or satellite SAR measurements and validated with airborne LiDAR. The derived parameters were applied both prospectively and retrospectively, yielding hydrologically significant results.

Key findings include:

1. Radar-based SWE estimates show low bias and high correlation with concurrent LiDAR and SAR data.
2. The approach demonstrates potential for improved quantification of IWC flux and its translation to ground-level SWE.
3. The method offers significant implications for water resource management, potentially enhancing streamflow predictions and water allocation decisions.

This research, supported by a National Science Foundation Small Business Innovation Research grant, addresses critical hydrological challenges in the Western United States and has global implications for snow-dependent regions. Our presentation will address the methodology and results of the study.