AbstractPredictions of storm surge and flooding require models with higher resolution of coastal regions, to describe fine-scale bathymetric and topographic variations, natural and artificial channels, flow features, and barriers. However, models for real-time forecasting often use a lower resolution to improve efficiency. There is a need to understand how resolution of inland regions can translate to predictive accuracy, but previous studies have not considered differences between models that both represent conveyance into floodplains and are intended to be used in real time. In this study, the effects of model resolution and coverage are explored using comparisons between forecast-ready and production-grade models that both represent floodplains along the US southeast coast, but with typical resolutions in coastal regions of 400 and 50 m, respectively. For two storms that impacted the US southeast coast, it is shown that, although the overall error statistics are similar between simulations on the two meshes, the production-grade model allowed a greater conveyance into inland regions, which improved the tide and surge signals in small channels and increased the inundation volumes between 40% and 60%. Its extended coverage also removed water level errors of 20–40 cm associated with boundary effects in smaller regional models.