Effects of Moisture Initialization on MJO and its Teleconnection Prediction in BCC Subseasonal Coupled Model

It is well recognized that moisture is of great importance in the representation of the Madden-Julian Oscillation (MJO). In this study, a moisture initialization scheme is developed using the Beijing Climate Center (BCC) subseasonal operational coupled model, and its effects on MJO predictions are investigated. Three sets of hindcast experiments are conducted: a reference experiment without moisture initialization (REF) and two sensitivity experiments with moderate and strong moisture nudging schemes. We show that the prediction skill of the real-time multivariate MJO index is significantly improved in the two sensitivity experiments, where the skillful prediction can reach up to 20 days using the moderate nudging scheme. Larger improvements can be achieved when MJO was initiated from Phases 2-3 and 5-7 or during La Nina episodes, implying that the MJO prediction skill is more sensitive to moisture conditions in active convection regions and backgrounds. The moisture budget analysis indicated that the MJO prediction improvement using the moisture initialization schemes comes from the realistic representation of vertical moisture structure and vertical moisture advection to the east of the MJO convection in the planetary boundary layer. The improved MJO prediction is also demonstrated to enable a more reliable subseasonal prediction of extratropical circulation and precipitation through a more realistic description of MJO-related teleconnections. In addition, the ensemble mean of the three schemes can extend the MJO prediction skill to more than 22 days, showing the potential benefits of ensemble prediction of multiinitialization schemes.