Skip to main content
U.S. flag

An official website of the United States government

How do River DroughtCast models work, and how is uncertainty estimated?

We use machine learning models to forecast streamflow drought by training them on 20- and 40-year periods of past streamflow, weather, and watershed data. For a summary of our methods and how we evaluate the models, see:

  • Model input data, including daily weather, snow, soil moisture, and meteorological forecast, is summarized in our ScienceBase Catalog dataset.

Updated Date: March 31, 2026
Was this page helpful?