||A comparative analysis of two sets of downscaled simulations of the current climate and the future climate projections over Peninsular Florida (PF) and the West Florida Shelf (WFS) is presented to isolate the role of high-resolution air-sea coupling. In addition, the downscaled integrations are also compared with the much coarser, driving global model projection to examine the impact of grid resolution of the models. The WFS region is habitat for significant marine resources, which has both commercial and recreational value. Additionally, the hydroclimatic features of the WFS and PF contrast each other. For example, the seasonal cycle of surface evaporation in these two regions are opposite in phase to one another. In this study, we downscale the Community Climate System Model version 4 (CCSM4) simulations of the late twentieth century and the mid-twenty-first century (with reference concentration pathway 8.5 emission scenario) using an atmosphere only Regional Spectral Model (RSM) at 10 km grid resolution. In another set, we downscale the same set of CCSM4 simulations using the coupled RSM-Regional Ocean Model System (RSMROMS) at 10 km grid resolution. The comparison of the twentieth century simulations suggest significant changes to the SST simulation over WFS from RSMROMS relative to CCSM4, with the former reducing the systematic errors of the seasonal mean SST over all seasons except in the boreal summer season. It may be noted that owing to the coarse resolution of CCSM4, the comparatively shallow bathymetry of the WFS and the sharp coastline along PF is poorly defined, which is significantly rectified at 10 km grid spacing in RSMROMS. The seasonal hydroclimate over PF and the WFS in the twentieth century simulation show significant bias in all three models with CCSM4 showing the least for a majority of the seasons, except in the wet June-July-August (JJA) season. In the JJA season, the errors of the surface hydroclimate over PF is the least in RSMROMS. The systematic errors of surface precipitation and evaporation are more comparable between the simulations of CCSM4 and RSMROMS, while they differ the most in moisture flux convergence. However, there is considerable improvement in RSMROMS compared to RSM simulations in terms of the seasonal bias of the hydroclimate over WFS and PF in all seasons of the year. This suggests the potential rectification impact of air-sea coupling on dynamic downscaling of CCSM4 twentieth century simulations. In terms of the climate projection in the decades of 2041-2060, the RSMROMS simulation indicate significant drying of the wet season over PF compared to moderate drying in CCSM4 and insignificant changes in the RSM projection. This contrasting projection is also associated with projected warming of SSTs along the WFS in RSMROMS as opposed to warming patterns of SST that is more zonal and across the WFS in CCSM4.