The Role of Boundary Conditions in AMIP-2 Simulations of the NAO

The simulated North Atlantic Oscillation (NAO) teleconnection patterns and their interannual variability are evaluated from a suite of atmospheric models participating in the second phase of the Atmospheric Model Intercomparison Project (AMIP-2). In general the models simulate the observed spatial pattern well, although there are important differences among models. The NAO response to interannual variations in sea surface temperature (SST) and snow-cover boundary forcings are also evaluated. The simulated NAO indices are not correlated with the observed NAO index, despite being forced with observed SSTs, indicating that SSTs are not driving NAO variability in the models. Similarly, although a number of studies have identified a link between Eurasian snow extent and the phase of the NAO, no such link is apparent in the AMIP-2 results. It appears that, within the framework of the AMIP-2 experiments, the NAO is an internal mode of atmospheric variability and that impacts of SSTs and Eurasian snow cover on the phase of the NAO are not discernable. However, these conclusions do not necessarily apply to decadal-scale and longer variability or to coupled atmosphere–ocean models.