Landfalling tropical cyclones (TC) pose great threats to public safety. The recent decades have witnessed major advances of knowledge in TC dynamics and improvement in TC forecast models, however, occasionally inaccurate TC intensity and storm surge predictions remain a vital concern. Different representations of subgrid-scale physics by various atmospheric model parameterization schemes lead to uncertainty in predictions of TC’s intensity and associated surges. In a case study for Hurricane Arthur (2014), local closure scheme for planetary boundary layer turbulence produces lower equivalent potential temperature than non-local closure schemes, leading to under-predicted TC intensity and surge heights. On the other hand, higher-class cloud microphysics schemes over-predict TC intensity and surge heights. Without cumulus parameterization for coarse-resolution grids, both TC intensity and surge heights are grossly under-predicted due to large precipitation decreases in the storm center. To avoid widespread predictions, the ensemble mean approach is shown to be effective. Another source of TC forecast error is inaccurate sea surface temperature (SST) prediction, and accurate SST prediction necessitates a better understanding of mixing processes in the coastal ocean. Previously, the importance of TC-induced near-inertial currents (NICs) to mixing in the coastal ocean was overlooked. With high-frequency radar and autonomous glider, long-lasting NICs with amplitudes of ~0.4 m s-1 were observed on the shelf during Arthur. With an atmosphere-ocean model, we find the NICs were dominated by mode-1 vertical structure and were a major contributor to the shear spectrum. Therefore, NICs may be important in producing turbulent mixing and surface cooling during Arthur’s passage. In the future, with warmer SST, sea level rise, and possible hard shorelines in estuaries, increased storm surge hazard is expected. Using Isabel (2003) as a case study, we find storm intensification under 2100 SST raises surge heights in Chesapeake Bay by 0.1-0.4 m given increased energy input. While sea level rise in 2100 reduces surge heights by 0-0.15 m through non-linear processes, it increases total water level by 0.4-1 m. Moreover, hard shoreline further increases surge heights by up to 0.5 m in the middle and upper Chesapeake Bay by prohibiting energy flux towards wetlands.