Atmospheric Response to SST Anomalies: Seasonal and Background-State Dependence

One potential source of seasonal predictability for Europe is through the impact of SST (sea-surface temperature) anomalies on the atmosphere. However, the atmospheric response to SST anomalies is notoriously difficult to simulate, and may be sensitive to model details and biases, particularly biases in climatological jet position and in patterns of internal variability. The response of models to anomalous SST, and the dependence of the response on the background wind field, has often been studied by comparing the responses in different winter months, or by comparing responses across different models. In this work, we look at how an idealised GCM’s response to SST anomalies depends on the background wind using two approaches. The first approach is to configure the model to have an Earth-like arrangement of land and topography, and compare the model’s response to an SST anomaly in winter with its response in summer. In this approach the different seasons provide different background wind fields and patterns of internal variability. The second approach is to compare the model’s response to an SST anomaly in one season, but in three different configurations of the model. The three configurations used are an Earth-like arrangement of land and topography, a simplified Earth-like arrangement of land and topography, and an aquaplanet with no land or topography. By using these three configurations, we can study the how the response depends on the background wind field, and patterns of variability, within a given season. Our results indicate that the model responds very differently to SST anomalies in winter and summer, with summer responses tending to be linear and baroclinic, and winter responses projecting onto modes of internal variability. The details of both the summer and winter responses are different in the Atlantic and Pacific sectors, however. Reasons for these differences, and implications for seasonal predictability, will be discussed.