Modulation of Tropical Intraseasonal Oscillations by Ocean–Atmosphere Coupling

Abstract

The impact of ocean?atmosphere coupling on the structure and propagation characteristics of 30?60-day tropical intraseasonal oscillations (TISOs) is investigated by analyzing long-term simulations of the Meteorological Research Institute coupled general circulation model (CGCM) and its stand-alone atmospheric general circulation model (AGCM) version forced with SSTs derived from the CGCM and comparing them with recent observation datasets [Global Precipitation Climatology Project (GPCP) precipitation, 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40), and Reynolds SST]. Composite events of (i) eastward propagating Madden?Julian oscillations (MJOs) during boreal winter and (ii) northward propagating intraseasonal oscillations (NPISOs) during boreal summer, constructed based on objective criteria, show that the three-dimensional structure, amplitude, and speed of propagation, and the phase relationship among surface fluxes, SST, and convection, are markedly improved in the CGCM simulation. Consistent with the frictional wave conditional instability of the second kind mechanism, successive development of low-level convergence to the east (north) of deep convection was found to be important for eastward (northward) propagation of MJO (NPISO). Complex interaction between large-scale dynamics and convection reveals the importance of atmospheric dynamics and suggests that they are intrinsic modes in the atmosphere where coupling is not essential for their existence. However, as in observations, realistic coupling in the CGCM is found to result in the evolution of TISOs as coupled modes through a coherent coupled feedback process. This acts as an amplifying mechanism for the existing propagating convective anomalies and plays an important modifying role toward a more realistic simulation of TISOs. In contrast, the simulated TISOs in its atmosphere-alone component lack many of the important features associated with their amplitude, phase, and life cycle. Thus, a realistic representation of the interaction between sea surface and the atmospheric boundary layer is crucial for a better simulation of TISOs.