Authors: Nadia Schreiber – Colorado School of Mines
Title: Leveraging Probabilistic Forecasting in NextGen: Evaluating GEFS-Based Streamflow Predictions Across Hydrologic Models
Presentation Type: Poster
Abstract: This study advances the Next Generation National Water Model (NextGen) by integrating probabilistic ensemble forecasts from the Global Ensemble Forecast System (GEFS) to improve streamflow prediction in flood-prone basins. GEFS provides ensemble-based meteorological inputs that capture forecast uncertainty, offering a probabilistic alternative to traditional deterministic forecasting. The study evaluates GEFS-driven simulations against multiple modeling approaches, including the Conceptual Functional Equivalent (CFE) model, the Noah-OWP Modular framework, and a Long Short-Term Memory (LSTM) neural network. A case study on a flood-impacted basin assesses model performance in terms of flood timing, magnitude, and overall forecast uncertainty. Results highlight the advantages of ensemble forecasting in improving predictive skill and informing uncertainty-aware decision-making. The findings demonstrate the potential for GEFS-based predictions to enhance NextGen’s operational capabilities, supporting more robust hydrological forecasting and flood risk management.