Allen et al. (2026) Characterizing Improvements in Ensemble Forecast Performance Over the Last Decade: A Retrospective Analysis of the Hydrologic Ensemble Forecast Service (HEFS)

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Geophysical Research Letters
• Year: 2026
• Date: 2026-04-05
• Authors: M. C. Allen, Scott Steinschneider
• DOI: 10.1029/2025gl119088

Research Groups

• California Nevada River Forecast Center (RFC)
• National Weather Service (NWS)'s Hydrologic Ensemble Forecast Service (HEFS)

Short Summary

This study retrospectively analyzes the performance of operational short- and medium-range streamflow forecasts from the California Nevada River Forecast Center (2014-2025) using a novel hierarchical Bayesian model, finding improved performance for moderate and high flow events, though improvements in ensemble spread attributes are weak.

Objective

• To retrospectively analyze and quantify regional trends in deterministic and probabilistic performance of operational short- and medium-range streamflow ensemble forecasts over time, particularly addressing probabilistic performance and cases with limited data.

Study Configuration

• Spatial Scale: 97 sites within the California Nevada River Forecast Center region.
• Temporal Scale: Forecasts issued between water years 2014–2025, with lead times ranging from 1 to 14 days.

Methodology and Data

• Models used: National Weather Service (NWS)'s Hydrologic Ensemble Forecast Service (HEFS) for generating forecasts; a novel and generalizable hierarchical Bayesian model developed for performance analysis and trend quantification.
• Data sources: Operational short- and medium-range streamflow ensemble forecasts issued by the California Nevada River Forecast Center.

Main Results

• Improved performance was observed for moderate and high flow streamflow events.
• These improvements are potentially linked to meteorological model upgrades and enhanced data assimilation.
• The degree of improvement is dependent on the specific performance metric and forecast lead time.
• Stronger improvement trends were found for deterministic performance at shorter lead times.
• Only weak evidence was found for improvements in the attributes of ensemble spread.

Contributions

• Provides a novel retrospective analysis of operational ensemble forecast performance improvement over time, specifically addressing probabilistic performance and data-limited scenarios.
• Introduces a novel and generalizable hierarchical Bayesian model designed to partially pool data across sites and quantify regional trends in both deterministic and probabilistic forecast performance.

Funding

• Not specified in the abstract.

Citation

@article{Allen2026Characterizing,
  author = {Allen, M. C. and Steinschneider, Scott},
  title = {Characterizing Improvements in Ensemble Forecast Performance Over the Last Decade: A Retrospective Analysis of the Hydrologic Ensemble Forecast Service (HEFS)},
  journal = {Geophysical Research Letters},
  year = {2026},
  doi = {10.1029/2025gl119088},
  url = {https://doi.org/10.1029/2025gl119088}
}