Estimating the Sensitivity of the Atmospheric Teleconnection Patterns to SST Anomalies Using a Linear Statistical Method

Abstract

he Pacific?North American (PNA) pattern and the North Atlantic Oscillation (NAO) are known to contain a tropical sea surface temperature (SST)-forced component. This study examines the sensitivity of the wintertime NAO and PNA to patterns of tropical SST anomalies using a linear statistical?dynamic method. The NAO index is sensitive to SST anomalies over the tropical Indian Ocean, the central Pacific Ocean, and the Caribbean Sea, and the PNA index is sensitive to SST anomalies over the tropical Pacific and Indian Oceans. The NAO and PNA patterns can be reproduced well by combining the linear operator with the consistent SST anomaly over the Indian Ocean and the Niño-4 regions, respectively, suggesting that these are the most efficient ocean basins that force the teleconnection patterns. During the period of 1950?2000, the NAO time series reconstructed by using SST anomalies over the Indian Ocean + Niño-4 region + Caribbean Sea or the Indian Ocean + Niño-4 region is significantly correlated with the observation. Using a cross-spectral analysis, the NAO index is coherent with the SST forcing over the Indian Ocean at a significant 3-yr period and a less significant 10-yr period, with the Indian Ocean SST leading by about a quarter phase. Unsurprisingly, the PNA index is most coherent with the Niño-4 SST at 4?5-yr periods. When compared with the observation, the NAO variability from the linear reconstruction is better reproduced than that of the coupled model, which is better than the Atmospheric Model Intercomparison Project (AMIP) run, while the PNA variability from the AMIP simulations is better than that of the reconstruction, which is better than the coupled model run.