Abstract
The ability of an atmospheric general circulation model to simulate the observed primary modes of intraseasonal variability in the Northern Hemisphere upper-tropospheric winds during boreal winter is examined. The model used is the Navy Operational Global Atmospheric Prediction System. The authors examine differences between the observed and modeled modes of variability in the context of various model deficiencies, where the observed modes are derived from the European Centre for Medium-Range Weather Forecasts analyses. Rotated empirical orthogonal function analysis is used to determine the primary modes of variability in the Pacific and Atlantic regions. EOFs are computed for both the zonal and meridional wind components. Time-lagged composite analysis is used to examine the temporal evolution of these modes, as well as their relationship to tropical convection. Wave activity flux vectors are used to examine further the characteristics of these intraseasonal modes and their relationship to tropical and extratropical forcing.
It is found that the model simulates the extratropically forced modes well but simulates modes associated with tropical heating poorly. The poor tropical simulation is due primarily to the model's poor representation of the Madden–Julian oscillation (MJO). The model's inability to produce the MJO-related modes is reflected in the model upper-tropospheric wind field variability being too weak in the tropical and subtropical Indian and Pacific Ocean regions. Model biases in the simulated time-mean winds may also account for differences between the observed and modeled modes of variability.
