An intercomparison of approaches for improving operational seasonal streamflow forecasts

Summary

This intercomparison study evaluated multiple operational approaches for seasonal streamflow forecasting across five Pacific Northwest watersheds, systematically comparing statistical regression, ensemble streamflow prediction (ESP), climate-informed, and hybrid hierarchical methods. The authors found that forecast skill gains from climate information depend on watershed teleconnection strength and predictability characteristics; where teleconnections are weak, custom reanalysis-derived predictors outperformed standard climate indices. A new hierarchical ensemble streamflow prediction (HESP) approach provided robust, reliable forecasts across all initialization times and basin conditions.

UK applicability

UK water managers may benefit from the hierarchical ensemble approach and the insight that standard climate indices require watershed-specific customisation; however, direct applicability is limited as the study focuses on Pacific Northwest hydrology and teleconnection patterns specific to that region. Adaptation of methodology to UK precipitation-runoff relationships and Atlantic/North Atlantic Oscillation signals would be required.

Key measures

Forecast skill (correlation), bias, reliability of seasonal streamflow predictions across lead times; performance of statistical regression, ensemble streamflow prediction (ESP), climate-based, and hybrid approaches

Outcomes reported

The study compared multiple approaches to seasonal water supply forecasting across five Pacific Northwest watersheds, evaluating forecast skill, bias, and reliability across different lead times. The research assessed marginal benefits of initial hydrologic conditions, climate information, and hybrid methods for improving operational seasonal predictions.

Topic tags