Parametric Wind Forcing for Storm Surge Simulations Using the Unified Forecast System for Coastal Applications (UFS-Coastal)

To mitigate the impacts of extreme storm events (e.g., hurricanes), local officials need to have “quick” access to “accurate” predictions/forecasts of total water levels and flood inundation in preparation of flood protection and evacuation. These predictions are provided by storm surge modeling systems that include an atmospheric model that forces both a storm surge and a wave model, all coupled together, to produce reliable total water level forecasts. In this study, we use the ESMF/NUOPC based coupling model application UFS-Coastal as our storm surge modeling system (https://github.com/oceanmodeling/ufs-coastal/tree/feature/coastal_app) that utilizes the Community Data Models for Earth Prediction Systems (CDEPS) as its data component framework and the NUOPC based Community Mediator for Earth Prediction Systems (CMEPS) as its model coupling framework. In this exercise, the ocean model we use is SCHISM and the wave model is WAVEWATCH III (WW3), both forced by the Parametric Hurricane Modeling System (PaHM).
PaHM (https://github.com/noaa-ocs-modeling/PaHM) contains various light-weight parametric tropical cyclone (TC) models that require minimal computational resources to produce the wind fields on the fly and fast. Such models use limited physics in producing “accurate” wind fields in the vicinity of the storm’s path. The TC models in PaHM are complemented by an atmospheric boundary layer (ABL) model to improve its prediction capabilities. In UFS-Coastal, PaHM is coupled via CMEPS (one-way) with the ocean and the wave models while, SCHISM is coupled two-way with WW3 via the CMEPS mediator to generate the storm surge simulations for Hurricanes Sandy (2012) and Florence (2018) on the Eastern Coast of the United States. The storm surge simulations for Hurricanes Laura (2020) and Marco (2020) in the Gulf of Mexico, are combined into one by utilizing PaHM’s ability to generate the wind fields for multiple storms (interacting or not) in the same basin.
The simulation results were compared with observations (e.g., NOAA CO-OPS stations) to verify the integrity of the interpolation schemes between the coupled components and to validate the model outputs. Furthermore, the performance of the different TC models in PaHM was evaluated via comparisons with wind observations. The results produced under the UFS-Coastal framework for Hurricanes Sandy, Florence, Laura and Marco show promising performance on predicting both the total water levels and the flood inundation. Standard statistical measures were used to validate the performance of the coupled system and to identify spatial and temporal system limitations.