Representation of the Madden-Julian Oscillation: a tough test for a coupled model

The Madden-Julian Oscillation (MJO) represents the major mode of variability in tropical convection on timescales of 30-60 days. The MJO organizes convection on a regional scale in the tropics and by modulation of the strength and location of the main tropical heat source, it can generate teleconnection patterns which affect the weather in the sub- and extra-tropics. Associated with the organized convection are intense low-level wind anomalies which can generate oceanic Kelvin waves in the Pacific. These waves can have a significant impact on the thermal structure of the upper tropical Pacific and may have an impact on the ENSO cycle. For all these reasons, a good representation of the MJO is desirable in numerical models being used for atmospheric prediction on timescales of a few days up to several seasons, and in models being used to represent the global climate system. However, many studies have shown that the MJO is often rather poorly represented in numerical models. This study will discuss some aspects of representing the MJO, both in atmospheric GCMs forced with observed SSTs, and in fully coupled ocean-atmosphere GCMs. Because of the complex interaction between convection, large scale dynamics and the surface layers of the ocean it is clear that an accurate representation of the MJO provides one of the toughest tests for climate and forecast models.