Machine Learning Seasonal Precipitation Forecast for November 2025–January 2026

November 10, 2025

The Center for Western Weather and Water Extremes (CW3E) has issued the latest Machine Learning Seasonal Precipitation Forecast for the upcoming winter season (November 2025–January 2026). This forecast product provides probabilistic outlooks of seasonal precipitation anomalies across the western United States using an integrated framework that combines machine learning models (Random Forest, XGBoost, LSTM, and neural networks) with dynamical ensemble predictions from three North American Multi-Model Ensemble (NMME) systems: CCSM4, CFSv2, and GEOS. Specifically, a plurality of combined ML and NMME ensemble members tilt the odds towards drier than normal conditions across the southwestern U.S. and wetter than normal conditions across the northwestern U.S., with 50% ensemble agreement. ML and NMME forecasts are predicting patterns consistent with drier than normal conditions in Southern CA (4/7 members). A majority of the ML and NMME forecasts predict near-normal precipitation over Northern and Central CA (6/7 members).

This seasonal prediction system, based on the method introduced in Gibson et al. (2021), aims to provide situational awareness guidance to water resource managers in California and throughout the western U.S. region. Visit the S&S forecast webpage to explore the latest results. This effort is part of CW3E’s ongoing mission to advance research-to-operations forecasting capabilities and strengthen climate-resilient decision support for the western United States.

Summary provided by Z. Yang and M. J. DeFlorio. Forecast generated by B. Kawzenuk and Z. Yang.

Gibson, P. B., W. E. Chapman, A. Altinok, L. Delle Monache, M. J. DeFlorio, and D. E. Waliser (2021), Training machine learning models on climate model output yields skillful interpretable seasonal precipitation forecasts. Nature Communications Earth & Environment, 2, 159. https://doi.org/10.1038/s43247-021-00225-4.