Abstract: Extreme sea levels can cause coastal flooding and have devastating impacts on coastal societies. Extreme sea levels from storm tide (storm surge plus tide) are often sufficiently modeled by two-dimensional (2-D) barotropic shallow water equations, which allows us to generate computationally efficient global coastal flood models with variable resolution — potentially as fine as ~10-100 m in coastal regions of concern. Nevertheless, ongoing sea level rise and higher-frequency fluctuations in mean sea level from seasonal heating-cooling, changes in ocean current flow rates, and mesoscale eddies lead to uncertainty in forecasting and projecting coastal flooding. These effects are often accounted for in ad hoc ways by adding on a mean sea level offset before a forecast and adding increments of sea level rise for projections. However, recent developments into a global 2-D storm tide modeling system (ADCIRC-based model) show that the fluctuations in the mean sea level can be reasonably accounted for by dynamically incorporating the internal density structure from an ocean general circulation model, which has coarse coastal resolution. Moreover, I will show that this approach improves coastal sea level fluctuations across a broad frequency spectrum compared to the barotropic storm tide model, increasing potential forecasting skill, and that the information from the ocean general circulation model is accurately downscaled to the coast, which could facilitate dynamic high-resolution coastal flooding projections using inputs from atmosphere and ocean general circulation models.