Why
Accurate seasonal water supply forecasts are crucial for effective water resources management in the Western United States. This region faces dry conditions and high demand for water, and these forecasts are essential for making informed decisions. They guide everything from water supply management and flood control to hydropower generation and environmental objectives.
Solution
This challenge sought models for creating probabilistic forecasts of naturalized cumulative streamflow volume for 26 sites in the Western U.S. over specific seasonal periods. Participants explored over two dozen data sources, including antecedent streamflow, snowpack conditions, meteorological conditions, seasonal temperature and precipitation forecasts, climate teleconnection indices, and more.
Models were evaluated over multiple challenge stages based on forecast accuracy and additional factors such as model rigor, innovation, generalizability, efficiency, scalability, and report clarity. The stages included near-real time forecasts during the 2024 season as well as a rigorous cross-validation over historical streamflow data. Panels of hydrologists, water resource experts, and machine learning practitioners from U.S. federal agencies judged the submissions. Additionally, an Explainability and Communication Bonus Track challenged participants to create clear and useful forecast summaries for key forecast consumers like water resource managers.
| Category | Prize Pool |
|---|---|
| Overall | $325,000 |
| Hindcast Stage | $50,000 |
| Forecast Stage | $50,000 |
| Explainability and Communication Bonus Track | $75,000 |
| Total | $500,000 |
Results
Over the course of the challenge, 64 teams submitted predictions and other materials. Congratulations to the winning teams:
| Award | Overall | Hindcast Stage | Forecast Stage | Explainability Bonus |
|---|---|---|---|---|
| 1st Place | oshbocker | rasyidstat | oshbocker | kurisu |
| 2nd Place | ck-ua | ck-ua | ck-ua | kamarain |
| 3rd Place | rasyidstat | oshbocker | rasyidstat | rasyidstat |
| 4th Place | kurisu | oshbocker | ||
| 5th Place | progin | iamo-team |
In the Final Stage, 20 teams submitted cross-validation predictions for this challenge's most statistically robust evaluation of models' forecast skill. Error scores from the cross-validation predictions are shown in the plot below.
See the final results announcement for more information on the winning approaches. All of the prize-winning solutions and model reports from this competition, along with the datasets assembled for the challenge, are linked below and available for anyone to use and learn from. The dataset from this challenge is also available for ongoing learning and development. The data is free to use outside of the challenge with attribution.
The competitions
Forecast Stage
Water managers in the Western U.S. rely on accurate water supply forecasts to better operate facilities and mitigate drought. Help the Bureau of Reclamation improve seasonal water supply estimates in this probabilistic forecasting challenge! [Forecast Stage] #climate
Final Prize Stage
Water managers in the Western U.S. rely on accurate water supply forecasts to better operate facilities and mitigate drought. Help the Bureau of Reclamation improve seasonal water supply estimates in this probabilistic forecasting challenge! [Final Prize Stage] #climate
Hindcast Evaluation Arena
Water managers in the Western U.S. rely on accurate water supply forecasts to better operate facilities and mitigate drought. Help the Bureau of Reclamation improve seasonal water supply estimates in this probabilistic forecasting challenge! [Hindcast Evaluation Arena] #climate
Hindcast Development Arena
START HERE! Water managers in the Western U.S. rely on accurate water supply forecasts to better operate facilities and mitigate drought. Help the Bureau of Reclamation improve seasonal water supply estimates in this probabilistic forecasting challenge! #climate