New study examines performance of the National Water Model during an intense, localized flash flood event
In late May 2018, Ellicott City, Maryland experienced catastrophic flash-flooding. NOAA’s National Water Model (NWM) was developed to simulate observed and forecast streamflow over the entire continental United States. A new study led by CIRES and NOAA researchers at the Physical Sciences Laboratory examines performance of the NWM during the intense, localized event at Ellicott City. The study uses multiple evaluation techniques to investigate various sources of error from both the precipitation forecasts themselves, and errors that may be due to the hydrologic model component. The results, recently published in the Journal of Hydrometeorology, show that this “multiscale” method can guide both ongoing research and model improvement to ultimately advance flash flood prediction.
The study found that for this specific flash flood event, the operational meteorological forecast was reasonably skillful, and the NWM streamflow forecasts in the small basin above Ellicott City suggested the potential for flash flooding. In larger surrounding basins, the NWM streamflow response was more complex, and forecast errors were found to be influenced by both hydrologic process representation in the NWM, as well as meteorological precipitation forecast errors.
The NWM was operationally implemented in 2016 to provide forecasts and analyses of streamflow for 2.7 million river reaches and other hydrologic information over the entire United States, expanding to previously underserved locations. Considering the great potential of the model, its relatively young implementation, and the short history of published results, this study is an important early step to fulfilling the shared community need to evaluate the NWM across various regions and a diverse suite of hydrologic conditions and flood events.
The article "A Multiscale, Hydrometeorological Forecast Evaluation of National Water Model Forecasts of the May 2018 Ellicott City, Maryland, Flood" is by Francesca Viterbo, Kelly Mahoney, Laura Read, Fernando Salas, Bradfor Bates, Jason Elliott, Brian Cosgrove, Aubrey Dugger, David Gochis, and Robert Cifelli: https://doi.org/10.1175/JHM-D-19-0125.1.
Posted: May 4, 2020