Evolving Coastal Total Water Level Models Referenced to the Geoid across the Globe from Laplace to Today (Core Science Keynote)

The foundations of both ocean and atmospheric simulation models date back to 1775 when Laplace published the Laplace Tidal Equations that were later generalized to the Shallow Water Equations. Laplace laid the foundation of modeling tides at a global scale. Computational models of the coastal ocean and its adjacent floodplains emerged with the advent of digital computing in the late 1960’s and have especially evolved over the past two decades along with computational power, ever more efficient and accurate algorithms and vast improvements in observational data. The key factors in coastal water level modeling skill are the underlying process physics, the algorithms applied to form the discrete analog of the governing partial differential equations, the mesh resolution of the processes and physical system, and the efficiency of the algorithms and code. This talk will examine these factors and how they have evolved in NOAA’s Storm and Tide Operational Forecast System (STOFS) 2D Global.

Tides in STOFS 2D Global are forced with direct total gravitational pull forces and dissipated using a directional internal tide generation/dissipation formulation and highly localized bottom boundary layer dissipation, Atmospheric forcing includes nested GFS-FV3 and HRRR winds with nudging algorithms to improve core hurricane wind speed intensity. The CICE model provides ice coverage, controlling both air-ice-sea momentum transfer and wind wave propagation/dissipation. Downscaled temperature and salinity fields from Global RTOFS drive the baroclinic pressure gradient terms and the earth’s thermohaline circulation with vital features such as the Gulf Stream pulling away on the order of a meter of water from the U.S. East Coast. Coupling to the National Water Model drives hydrology in rivers and across the floodplain allowing a dynamic connection between upland hydrology and the ocean. Finally coupling to wind wave models such as Wavewatch III and SWAN allow for wave radiation stresses to drive coastal set up. All these physics demand that both the physical system, i.e. geometry, bathymetry, and topography, be well resolved and that the driving algorithms do not add artificial dissipation that overwhelms the actual physical dissipation in the system. Finally code efficiency is paramount to accuracy since mesh resolution ultimately controls model skill in representing the myriad of physical processes.