- Reclamation
- R&D
- Research Projects
- Machine Learning for Improving Sub-Seasonal Forecasting
Machine Learning for Improving Sub-Seasonal Forecasting
Project ID: 20204
Principal Investigator: Kenneth Nowak
Research Topic: Water Supply Forecasting
Funded Fiscal Years:
2020
Keywords: None
Research Question
Reclamation concluded the prize competition "Sub-Seasonal Climate Forecast Rodeo" in June 2019 with a symposium hosted at NOAA headquarters in Silver Spring, MD. Several winning teams that were able to outperform operational forecasts from NOAA used machine learning techniques to produce their forecasts. This funding would allow Reclamation to partner with those teams or pursue refinement of their solutions by other means. In addition to improving sub-seasonal forecast skill, Reclamation will be able to build and enhance internal machine learning capacity.
Need and Benefit
Furthermore, Reclamation has launched a second forecast rodeo (Rodeo II), hosted by TopCoder. The Rodeo II hindcast phase (warm-up prior to another year-long competition) was completed August 2019 and winning solutions have been transferred to Reclamation. Funding for this activity would facilitate evaluation of Rodeo II hindcast winning methods. This work supports the Federal Action Plan for Improving Forecasts of Water Availability, prepared pursuant to Section 3 of the Presidential Memorandum.
Contributing Partners
Contact the Principal Investigator for information about partners.
Research Products
Please contact research@usbr.gov about research products related to this project.
