Evaluation of National Water Model Streamflow Forecasts for Large- and Small- Scale Flood Events during the 2024 Hurricane Season

One of the major objectives of high-resolution, continental-scale streamflow forecasts, such as the NOAA National Water Model (NWM) is to provide useful guidance for emergency managers, reservoir operators and a range of other stakeholders during high flow or flood events. To assess the extent to which the forecasts achieve, or progress toward, this objective, it is essential to evaluate the quality of the operational forecasts that were generated prior to and during flood events. For forecast end-users, post-event forecast evaluations provide insight about the accuracy and utility of the forecasts as manifested during a real event, which can help these groups to better understand and make better use of the forecasts for decision-making during future events. For forecasting agencies, post-event evaluations can highlight issues related to forcing inputs, hydrologic model formulation, data assimilation techniques, or other aspects of the forecast process that warrant further investigation and that might only become apparent during operations. Across groups, post-event hydrologic forecast evaluations provide real, concrete examples of the potential utility, challenges and sources of error associated with continental-scale hydrologic forecasting, which can in turn foster more informed discussions, more informed use by stakeholders, and evidence-driven prioritization of improvements.

While post-event evaluations are important, the magnitude of the data associated with high-resolution forecasts, particularly for large flood events, present practical challenges for performing and synthesizing the results of the evaluation. The NOAA-National Water Center has been developing protocols, methods and tools to evaluate NWM streamflow forecasts following flood events. These tools and methods were tested during the 2024 hurricane season. This presentation aims to share evaluation results from both large-scale (e.g., Hurricanes Beryl and Debby) and small-scale events (e.g., post-fire floods in Ruidoso, NM). Key findings related to NWM forecast performance for each event will be highlighted, along with a discussion of the impact of event scale on the methods, challenges, and findings of single-event streamflow forecast evaluations.