A High-Resolution National-Scale Hydrologic Forecast System from a Global Ensemble Land Surface Model

Area of Science:

• Hydrology
• Environmental Science
• Computer Science

Background:

• Existing large-scale streamflow prediction systems operate at coarse resolutions, limiting their effectiveness for localized flood warnings.
• Millions of people could benefit from advanced flood prediction systems capable of several-day advance warnings.

Purpose of the Study:

• To develop a method for routing global runoff ensemble forecasts to produce high-resolution, 15-day streamflow predictions.
• To create approximate recurrence intervals and warning points for exceeding streamflow thresholds.
• To enhance national flood forecasting capabilities.

Main Methods:

• Utilized the Routing Application for Parallel computatIon of Discharge (RAPID) to route global runoff ensemble forecasts and historical runoff from the European Centre for Medium-Range Weather Forecasts (ECMWF) model.
• Employed distributed computing on computer clusters to process large watersheds with high-density stream networks.
• Developed the Streamflow Prediction Tool web application for visualizing results at regional and reach levels.

Main Results:

• Generated high spatial resolution 15-day stream forecasts, approximate recurrence intervals, and warning points.
• Enabled processing of large watersheds and high-density stream networks through distributed computation.
• Integrated ensemble predictions for nearly 2.7 million reaches of the National Hydrography Plus Version 2 Dataset.

Conclusions:

• The developed method and tool can significantly enhance national flood forecasting ability.
• High-resolution streamflow predictions and warning points can be generated effectively for large-scale applications.
• The Streamflow Prediction Tool serves as a base for improved hydrologic forecasting services.