Potential Salinity and Temperature Futures for the Chesapeake Bay Using a Statistical Downscaling Spatial Disaggregation Framework

Estuaries are productive and ecologically important ecosystems, incorporating environmental drivers from water-sheds, rivers, and the coastal ocean. Climate change has potential to modify the physical properties of estuaries, with impacts on resident organisms. However, projections from general circulation models (GCMs) are generally too coarse to resolve important estuarine processes. Here, we statistically downscaled near-surface air temperature and precipitation projections to the scale of the Chesapeake Bay watershed and estuary. These variables were linked to Susquehanna River streamflow using a water balance model and finally to spatially resolved Chesapeake Bay surface temperature and salinity using statistical model trees. The low computational cost of this approach allowed rapid assessment of projected changes from four GCMs spanning a range of potential futures under a high CO₂ emission scenario, for four different downscaling methods. Choice of GCM contributed strongly to the spread in projections, but choice of downscaling method was also influential in the warmest models. Models projected a ~2-5.5 °C increase in surface water temperatures in the Chesapeake Bay by the end of the century. Projections of salinity were more uncertain and spatially complex. Models showing increases in winter-spring streamflow generated freshening in the Upper Bay and tributaries, while models with decreased streamflow produced salinity increases. Changes to the Chesapeake Bay environment have implications for fish and invertebrate habitats, as well as migration, spawning phenology, recruitment, and occurrence of pathogens. Our results underline a potentially expanded role of statistical downscaling to complement dynamical approaches in assessing climate change impacts in dynamically challenging estuaries.