Assessment of Operational and Retrospective NOAA National Water Model Performance for Low Frequency Flooding Events: National Synthesis and Case Studies

The NOAA National Water Model (NWM) is a continuous cycling, operational hydrologic prediction system that provides high resolution forecasts of streamflow and other hydrologic variables for the contiguous U.S. The model architecture behind the NWM is the community WRF-Hydro modeling system which offers multi-scale and multi-physics configuration options for representation of terrestrial hydrologic processes such as infiltration, snowpack, evapotranspiration, vadose zone transport overland flow and inundation, shallow groundwater flow, streamflow and a simple reservoir treatment. The operational NWM is implemented to provide real-time, hourly analyses as well as short (18 hrs), medium (10 days) and long (30 days) range forecasts, driven by operational radar precipitation data (from the NSSL Multi-Radar/Multi-Sensor product) and numerical weather prediction models. Since August of 2016, the NWM has provided real-time operational forecasts as well as multiple long-term retrospective simulations as guidance for official flood forecasting operations for the National Weather Service. In this presentation, we provide an assessment of both forecast and retrospective simulation performance in the representation of low frequency flooding events. Contingency based statistics using both stakeholder defined flood thresholds and statistically-defined flood thresholds are shown. There are clear differences in model flood representation capabilities that vary across the nation as functions of hydrologic regime and basin scale. In addition to nationally-synthesized flood evaluation results, a small set of notable, recent case studies are presented in order to highlight the different kinds of model information that is available for flood guidance.