Assessment of the Parametric Rainfall (P-CLIPER) Component within a Compound Coastal Flood Modeling System

Coastal communities are increasingly impacted by flooding due to multiple components of natural hazards: intense precipitation, riverine flooding and storm surge. With support from NOAA, NSF Coastlines and People (CoPe), and the NIST Center of Excellence on Risk-based Community Resilience Planning, a compound flood modeling system has been developed that couples storm tracking, wind field, precipitation, hydrologic, wave and hydrodynamic models to predict the “total water level” (tides + waves + surge + runoff) due to a storm event, either for current or future climate. However, the predictions of a compound system are only as good as the individual components. In this work, improvements are made to the parametric rainfall model (P-CLIPER), which uses the track location and wind speed to compute hourly averaged rainfall fields during a tropical cyclone event, and further assessment is completed for the calibration of the precipitation model. Rainfall fields generated by P-CLIPER can be used both for forcing of a hydrologic model and direct precipitation over the hydrodynamic model domain; each of which can significantly impact the coastal domain of interest. Improvements have been made to allow the storm category to vary with time during the P-CLIPER simulation, rather than remain fixed at the landfall storm category. Additionally, for ensemble modeling with the coupled system, the random assignment for the P-CLIPER rain intensity parameter is now being drawn from a truncated normal distribution to more closely mimic the underlying distribution used to fit the P-CLIPER model to satellite data. Recent work has suggested that the P-CLIPER model must be calibrated both for a geographic region and the specific hydrologic model that will be used in the compound flooding system. Calibration of P-CLIPER entails determining the range of acceptable rain intensity values used for individual or ensemble modeling applications in a particular geographic region. The P-CLIPER model was previously calibrated to the North Carolina coastal region primarily through the hydrologic response created from the simulated precipitation fields for twelve historic storms from 2002 through 2008 using Stage IV radar for comparison. This study assesses over twenty historical storms from 2002 through 2023 for the same region using Multi-Radar Multi-Sensor (MRMS) precipitation fields and United States Geological Survey (USGS) streamflow measurements for comparison. For this study, background is provided for the coupled model system and study area, analysis of the hydrologic response compared to measured USGS gage station data is presented, and analysis of the accumulated precipitation compared to MRMS radar measurements is presented for several historical events in the Tar and Neuse watersheds of North Carolina.