Estimating the Continental Response to Global Warming Using Pattern-Scaled Sea Surface Temperatures and Sea Ice

Better constraining the continental climate response to anthropogenic forcing is essential to improve climate projections. In this study, pattern scaling is used to extract, from observations, the patterned response of sea surface temperature (SST) and sea- ice concentration (SICE) to anthropogenically dominated long-term global warming. The SST response pattern includes a warming of the tropical Indian Ocean, the high northern latitudes, and the western boundary currents. The SICE pattern shows seasonal variations of the main locations of sea ice loss. These SST/SICE response patterns are used to drive an ensemble of the atmospheric general circulation model NCAR/CAM5 (National Center for Atmospheric Research Community Atmosphere Model 5) over the period 1980-2010, along with a standard "AMIP" ensemble using observed SST/SICE. The simulations enable attribution of a variety of observed trends of continental climate to global warming. On the one hand, the warming trends observed in all seasons across the entire Northern Hemisphere extratropics result from global warming, as well as the snow loss observed over the northern mid-latitudes and northwestern Eurasia. On the other hand, 1980-2010 precipitation trends observed in winter over North America and in summer over Africa result from the recent decreasing phase of the Pacific Decadal Oscillation and the recent increasing phase of the Atlantic Multidecadal Oscillation, respectively, which are not part of the global warming signal. The method holds promise for near-term decadal climate prediction, but as currently framed cannot distinguish regional signals associated with oceanic internal variability from aerosol forcing and other sources of short-term forcing.