Addressing SST Uncertainty for Future Precipitation Extremes within the Southeast and Caribbean US at Convective Permitting Scales

A well-documented problem within the CMIP5 ensemble is a cold SST bias when compared against observations within the North Atlantic. It is anticipated that this cold SST bias will have significant impacts on the rainfall intensity and duration. Global climate models also do not properly resolve some important ocean currents that are important for cyclogenesis and storm intensity, which will also impact precipitation extremes. This study investigates the impact of the cold CMIP5 SST bias and the ability to resolve important SST features for traditional dynamical downscaling, which does not include bias corrections, on precipitation extremes within the Southeast and Caribbean US. Sensitivity experiments are performed by dynamically downscaling a CMIP5 ensemble member at convective permitting scales for a historical and future time slice to address the SST uncertainties. The convective permitting simulations will be compared with prior dynamically downscaled simulations of the same CMIP5 ensemble member at coarser resolution (36-km) as well as additional convective permitting simulations both of which do not correct the cold SST bias nor properly resolve important SST features. This comparison will help document the added value of convective permitting simulations with improved SSTs for climate change projections for the Southeast and Caribbean US.