Interseasonal Prediction of Western United States Snowpack with Deep Learning

Gemma Anderson | 19-ERD-032

Executive Summary

During this project, we will train a deep neural network on 40 years of observational data supplemented with high-resolution model data to better understand the underlying physical processes of the snowpack in the Western United States. The resulting state-of-the-art deep-learning computational tools will improve snowpack predictions that look six to ten months in advance, enabling water resource managers to plan for the year ahead.

Publications, Presentations, and Patents

Anderson, G. J. 2020. "Probabilistic deep learning for seasonal forecasting." 2020 American Geophysical Union Fall Meeting (online), December 2020. Abstract. LLNL-ABS-812531

Goncalves, A. R., et al. 2019. "Inter-seasonal snowpack prediction with deep convolutional neural networks." 2019 American Geophysical Union Annual Meeting, San Francisco, CA, December 2019. LLNL-POST-798858

Lucas, D. D. 2020. "Changing Snow Properties Improves Simulation of Southern Hemisphere Sea Ice." Geophysical Research Letters. Abstract. LLNL-ABS-812770

Pan, B., et al., 2020a. "Identifying and correcting climate projection biases using artificial intelligence." 2020 American Geophysical Union Annual Meeting (online), December 2020. LLNL-ABS-812865

——— 2020b. "Improving seasonal forecast using probabilistic deep learning." LLNL-PRES-814574