Forcing of Northern Hemisphere climate trends

The impact of observed global SST trends during the second half of the twentieth century on the Northern Hemisphere extratropical winter atmospheric circulation is investigated using ensembles of simulations with the Center for Ocean - Land - Atmosphere Studies ( COLA) atmospheric GCM. In contrast to some other studies, the simulated ensemble mean 500-hPa trends in the North Atlantic sector do not resemble the observed trend. However, the intraensemble variability of the trends is large, with the dominant structure of that variability resembling the Arctic Oscillation "annular mode.'' The model results are consistent with the interpretation that the observed trend is dominated by the forced signal in the Pacific - North America sector, while over the rest of the Northern Hemisphere, and especially the North Atlantic sector, the trend is primarily interdecadal timescale internal atmospheric noise with an annular structure. In order to diagnose the origins of the forced component of the model trend, a series of equilibrium response simulations is performed using constant-in-time SST anomalies with the structure of the trend superimposed on the annually varying climatological SST. It is found that the SST trend in the latitude belt from 20degreesS to 20degreesN is responsible for forcing much of the extratropical trend, and that the dominant tropical forcing is the SST trend in the Indian Ocean/western Pacific and eastern Pacific sectors. The idealized experiments show that the precipitation response in the Tropics is linearly related to the SST trend, and that the NH December - January February height response to SST anomalies in various regions is quasi-linear. Some additional analysis and interpretation is given. The extratropical response to low-latitude SST trends in the idealized experiments has characteristics reminiscent of Rossby wave trains forced by tropical deep convection. The intraensemble variability in the model's extratropical zonal mean height trend, which cannot be explained by external forcing, appears to be due to variability in the trends of midlatitude eddy stirring. The observed zonal mean trend also shows evidence of forcing by trends in the eddy stirring.