Response of the GISS General Circulation Model to a Midlatitude Sea Surface Temperature Anomaly in the North Pacific

Abstract

The influence of a North Pacific sea surface temperature (SST) anomaly on the wintertime atmospheric circulation is investigated with the GISS general circulation model (GCM) 2. Although no signal could be detected by the standard univariate t-test, a multivariate statistical analysis based on the assumption that the atmospheric response is primarily at large scales shows that the SST anomaly has an influence on the model Northern Hemisphere climate. The signal, primarily barotropic, is strongest at zonal wavenumbers 3 to 5. It is above the noise level in the middle and upper troposphere, but not near the ground. For realistic SST magnitudes, the change in geopotential height could reach several tens of meters, suggesting that midlatitude SST anomalies may have a weak climatic impact. However, the signal is model-dependent since it differs from the response of the (less realistic) GISS model 1 to the same SST anomaly. The signal is also inconsistent with 500 mb height anomalies observed during two periods with similar SST anomalies in the North Pacific. A two-layer quasi-geostrophic linear model with a zonally symmetric basic state is then used to investigate whether the GCM response can he interpreted in terms of forced stationary waves. When the mean zonal flow and the anomaly heating field are prescribed from the GCM data, it is found that the linear model prediction is consistent with the GCM signal, although only a small fraction of the anomaly variance can be explained. A simple linear wave model is thus useful to analyze the GCM experiments, but it cannot be used for predictive purposes, unless the relation between SST and diabatic heating anomaly can be better established in the mid-latitudes.