Subgrid correction of storm surge modeling in orthogonal curvilinear coordinates

Ensemble forecasts of storm surge require rapid computations of storm surge models over a fixed cycle, usually with relatively coarse-grid models. This work describes numerical extensions to include subgrid approaches accounting for the bulk effect of high-resolution bathymetry and bottom roughness to increase the accuracy of these coarse-grid setups. In particular, the subgrid corrections are developed in a semi-implicit staggered-grid finite difference formulation for the 2D non-linear Shallow Water Equations expressed in an orthogonal curvilinear coordinate system. This general framework is implemented for the Coastal and Estuarine Storm Tide (CEST) model. Verification and validation using a number of test cases, ranging from an idealized test case with analytical solutions to real-world storm surge problems, are conducted in order to assess the performance of the subgrid features. Comparisons of solution accuracy and speed show the subgrid implementation consistently improves cost versus accuracy performance. • Semi-implicit formulations for 2D Shallow Water Equations with subgrid corrections. • Subgrid corrections enhance solution accuracy in orthogonal curvilinear grids. • Subgrid corrections reduce grid resolution needs, enhancing small channel connectivity. • Accuracy gains from subgrid corrections offset with modest computational cost increases.