Enhancing Deterministic Freezing-Level Predictions in the Northern Sierra Nevada through Deep Neural Networks

Abstract Accurate prediction of the freezing level (FZL) is essential for hydrometeorological forecasting systems, with direct implications for runoff generation and reservoir management. In this study, we develop a deep learning–based postprocessing framework using the U-Net convolutional neural network (CNN) architecture to refine the FZL forecasts from the West Weather Research and Forecasting (West-WRF) Model. The proposed framework leverages reforecast data from West-WRF and FZL estimates from the California Nevada River Forecast Center (CNRFC) to develop U-Net models over the northern Sierra Nevada watersheds, such as the hydrologically critical Yuba–Feather watershed. We introduce two U-Net model variants, U-Net-log and U-Net Gaussian mixture model (U-Net-GMM), that employ specialized loss functions beyond the standard benchmarks to enhance forecast skill. U-Net-log utilizes the log-cosh cosine of error, and U-Net-GMM uses Gaussian mixture model loss functions, to enhance FZL forecasts. Results show that U-Net-based postprocessing reduces centered root-mean-square errors by up to 20% and increases forecast–observation correlation by about 10% compared to raw West-WRF. Evaluation using the continuous ranked probability score (CRPS) for U-Net-GMM further demonstrates consistent improvements across lead times. While performance fluctuates with forecast horizon, storm variability, and diurnal forcing, U-Net-GMM and U-Net-log consistently outperform the baseline. The models capture the spatiotemporal variability of the FZL across different elevations, mitigating biases from the West-WRF Model. This novel deep learning–based postprocessing approach demonstrates a promising pathway for integrating machine learning into hydrometeorological forecasting and decision support within the Forecast-Informed Reservoir Operations (FIRO) framework.